Package 'tarchetypes'

Name of the target. tar_cue_age() expects an unevaluated symbol for the name argument, whereas tar_cue_age_raw() expects a character string for name.

R code to run the target and return a value.

A difftime object of length 1, such as as.difftime(3, units = "days"). If the target's output data files are older than age (according to the most recent time stamp over all the target's output files) then the target will rerun. On the other hand, if at least one data file is younger than Sys.time() - age, then the ordinary invalidation rules apply, and the target may or not rerun. If you want to force the target to run every 3 days, for example, set age = as.difftime(3, units = "days").

Code to define a dynamic branching branching for a target. In tar_target(), pattern is an unevaluated expression, e.g. tar_target(pattern = map(data)). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(pattern = quote(map(data))).

To demonstrate dynamic branching patterns, suppose we have a pipeline with numeric vector targets x and y. Then, tar_target(z, x + y, pattern = map(x, y)) implicitly defines branches of z that each compute x[1] + y[1], x[2] + y[2], and so on. See the user manual for details.

Logical, whether to enable tidy evaluation when interpreting command and pattern. If TRUE, you can use the "bang-bang" operator ⁠!!⁠ to programmatically insert the values of global objects.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Logical, whether to rerun the target if the user-specified storage format changed. The storage format is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the user-specified storage repository changed. The storage repository is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the user-specified iteration method changed. The iteration method is user-specified through tar_target() or tar_option_set().

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

A targets::tar_cue() object. (See the "Cue objects" section for background.) This cue object should contain any optional secondary invalidation rules, anything except the mode argument. mode will be automatically determined by the age argument of tar_age().

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

An expression with special syntax to define a collection of targets in a pipeline. Example: tar_assign(x <- tar_target(get_data())) is equivalent to list(tar_target(x, get_data())). The rules of the syntax are as follows:

• The code supplied to tar_assign() must be enclosed in curly braces beginning with ⁠{⁠ and ⁠}⁠ unless it only contains a one-line statement or uses = as the assignment.

• Each statement in the code block must be of the form x <- f(), or x = f() where x is the name of a target and f() is a function like tar_target() or tar_quarto() which accepts a name argument.

• The native pipe operator ⁠|>⁠ is allowed because it lazily evaluates its arguments and be converted into non-pipe syntax without evaluating the code.

Symbol, name of the target. In tar_target(), name is an unevaluated symbol, e.g. tar_target(name = data). In tar_target_raw(), name is a character string, e.g. tar_target_raw(name = "data").

A target name must be a valid name for a symbol in R, and it must not start with a dot. Subsequent targets can refer to this name symbolically to induce a dependency relationship: e.g. tar_target(downstream_target, f(upstream_target)) is a target named downstream_target which depends on a target upstream_target and a function f(). In addition, a target's name determines its random number generator seed. In this way, each target runs with a reproducible seed so someone else running the same pipeline should get the same results, and no two targets in the same pipeline share the same seed. (Even dynamic branches have different names and thus different seeds.) You can recover the seed of a completed target with tar_meta(your_target, seed) and run tar_seed_set() on the result to locally recreate the target's initial RNG state.

R code to run the target. In tar_target(), command is an unevaluated expression, e.g. tar_target(command = data). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(command = quote(data)).

R code for the upstream change-inducing target.

Whether to invoke tidy evaluation (e.g. the ⁠!!⁠ operator from rlang) as soon as the target is defined (before tar_make()). Applies to arguments command and change.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Optional storage format for the target's return value. With the exception of format = "file", each target gets a file in ⁠_targets/objects⁠, and each format is a different way to save and load this file. See the "Storage formats" section for a detailed list of possible data storage formats.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, name of the iteration mode of the target. Choices:

• "vector": branching happens with vctrs::vec_slice() and aggregation happens with vctrs::vec_c().

• "list", branching happens with ⁠[[]]⁠ and aggregation happens with list().

• "group": dplyr::group_by()-like functionality to branch over subsets of a non-dynamic data frame. For iteration = "group", the target must not by dynamic (the pattern argument of tar_target() must be left NULL). The target's return value must be a data frame with a special tar_group column of consecutive integers from 1 through the number of groups. Each integer designates a group, and a branch is created for each collection of rows in a group. See the tar_group() function to see how you can create the special tar_group column with dplyr::group_by().

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date. Only applies to the downstream target. The upstream target always runs.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Name of the new target. tar_combine() expects unevaluated expressions for the name, and command arguments, whereas tar_combine_raw() uses a character string for name and an evaluated expression object for command. See the examples for details.

One or more target objects or list of target objects. Lists can be arbitrarily nested, as in list().

R command to aggregate the targets. Must contain !!!.x where the arguments are to be inserted, where ⁠!!!⁠ is the unquote splice operator from rlang.

tar_combine() expects unevaluated expressions for the name, and command arguments, whereas tar_combine_raw() uses a character string for name and an evaluated expression object for command. See the examples for details.

Logical, whether to insert the names of the targets into the command when splicing.

Code to define a dynamic branching branching for a target. In tar_target(), pattern is an unevaluated expression, e.g. tar_target(pattern = map(data)). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(pattern = quote(map(data))).

To demonstrate dynamic branching patterns, suppose we have a pipeline with numeric vector targets x and y. Then, tar_target(z, x + y, pattern = map(x, y)) implicitly defines branches of z that each compute x[1] + y[1], x[2] + y[2], and so on. See the user manual for details.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Optional storage format for the target's return value. With the exception of format = "file", each target gets a file in ⁠_targets/objects⁠, and each format is a different way to save and load this file. See the "Storage formats" section for a detailed list of possible data storage formats.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, name of the iteration mode of the target. Choices:

• "vector": branching happens with vctrs::vec_slice() and aggregation happens with vctrs::vec_c().

• "list", branching happens with ⁠[[]]⁠ and aggregation happens with list().

• "group": dplyr::group_by()-like functionality to branch over subsets of a non-dynamic data frame. For iteration = "group", the target must not by dynamic (the pattern argument of tar_target() must be left NULL). The target's return value must be a data frame with a special tar_group column of consecutive integers from 1 through the number of groups. Each integer designates a group, and a branch is created for each collection of rows in a group. See the tar_group() function to see how you can create the special tar_group column with dplyr::group_by().

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Name of the target. tar_cue_age() expects an unevaluated symbol for the name argument, whereas tar_cue_age_raw() expects a character string for name.

A difftime object of length 1, such as as.difftime(3, units = "days"). If the target's output data files are older than age (according to the most recent time stamp over all the target's output files) then the target will rerun. On the other hand, if at least one data file is younger than Sys.time() - age, then the ordinary invalidation rules apply, and the target may or not rerun. If you want to force the target to run every 3 days, for example, set age = as.difftime(3, units = "days").

Logical, whether to rerun the target if command changed since last time.

Logical, whether to rerun the target if the value of one of the dependencies changed.

Logical, whether to rerun the target if the user-specified storage format changed. The storage format is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the user-specified storage repository changed. The storage repository is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the user-specified iteration method changed. The iteration method is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the file(s) with the return value changed or at least one is missing.

Logical vector evaluated locally when the target is defined. If any element of condition is TRUE, the target will definitely rerun when the pipeline runs. Otherwise, the target may or may not rerun, depending on the other invalidation rules. condition is evaluated when this cue factory is called, so the condition cannot depend on upstream targets, and it should be quick to calculate.

Logical, whether to rerun the target if command changed since last time.

Logical, whether to rerun the target if the value of one of the dependencies changed.

Logical, whether to rerun the target if the user-specified storage format changed. The storage format is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the user-specified storage repository changed. The storage repository is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the user-specified iteration method changed. The iteration method is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the file(s) with the return value changed or at least one is missing.

Logical vector evaluated locally when the target is defined. If any element of condition is TRUE, the pipeline will skip the target unless the target has never been built before. If all elements of condition are FALSE, then the target may or may not rerun, depending on the other invalidation rules. condition is evaluated when this cue factory is called, so the condition cannot depend on upstream targets, and it should be quick to calculate.

Logical, whether to rerun the target if command changed since last time.

Logical, whether to rerun the target if the value of one of the dependencies changed.

Logical, whether to rerun the target if the user-specified storage format changed. The storage format is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the user-specified storage repository changed. The storage repository is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the user-specified iteration method changed. The iteration method is user-specified through tar_target() or tar_option_set().

Logical, whether to rerun the target if the file(s) with the return value changed or at least one is missing.

Symbol, name of the target. In tar_target(), name is an unevaluated symbol, e.g. tar_target(name = data). In tar_target_raw(), name is a character string, e.g. tar_target_raw(name = "data").

A target name must be a valid name for a symbol in R, and it must not start with a dot. Subsequent targets can refer to this name symbolically to induce a dependency relationship: e.g. tar_target(downstream_target, f(upstream_target)) is a target named downstream_target which depends on a target upstream_target and a function f(). In addition, a target's name determines its random number generator seed. In this way, each target runs with a reproducible seed so someone else running the same pipeline should get the same results, and no two targets in the same pipeline share the same seed. (Even dynamic branches have different names and thus different seeds.) You can recover the seed of a completed target with tar_meta(your_target, seed) and run tar_seed_set() on the result to locally recreate the target's initial RNG state.

Character vector of URLs to track and download. Must be known and declared before the pipeline runs.

Character vector of local file paths to download each of the URLs. Must be known and declared before the pipeline runs.

Method to be used for downloading files. Current download methods are "internal", "libcurl", "wget", "curl" and "wininet" (Windows only), and there is a value "auto": see ‘Details’ and ‘Note’.

The method can also be set through the option "download.file.method": see options().

If TRUE, suppress status messages (if any), and the progress bar.

character. The mode with which to write the file. Useful values are "w", "wb" (binary), "a" (append) and "ab". Not used for methods "wget" and "curl". See also ‘Details’, notably about using "wb" for Windows.

logical. Is a server-side cached value acceptable?

character vector of additional command-line arguments for the "wget" and "curl" methods.

named character vector of additional HTTP headers to use in HTTP[S] requests. It is ignored for non-HTTP[S] URLs. The User-Agent header taken from the HTTPUserAgent option (see options) is automatically used as the first header.

Character of length 1, name of the iteration mode of the target. Choices:

• "vector": branching happens with vctrs::vec_slice() and aggregation happens with vctrs::vec_c().

• "list", branching happens with ⁠[[]]⁠ and aggregation happens with list().

• "group": dplyr::group_by()-like functionality to branch over subsets of a non-dynamic data frame. For iteration = "group", the target must not by dynamic (the pattern argument of tar_target() must be left NULL). The target's return value must be a data frame with a special tar_group column of consecutive integers from 1 through the number of groups. Each integer designates a group, and a branch is created for each collection of rows in a group. See the tar_group() function to see how you can create the special tar_group column with dplyr::group_by().

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Starting expression. Values are iteratively substituted in place of symbols in expr to create each new expression, and then each new expression is evaluated.

tar_eval() expects an unevaluated expression for the expr object, whereas tar_eval_raw() expects an evaluated expression object.

List of values to substitute into expr to create the expressions. All elements of values must have the same length.

Environment in which to evaluate the new expressions.

Symbol, name of the target. In tar_target(), name is an unevaluated symbol, e.g. tar_target(name = data). In tar_target_raw(), name is a character string, e.g. tar_target_raw(name = "data").

A target name must be a valid name for a symbol in R, and it must not start with a dot. Subsequent targets can refer to this name symbolically to induce a dependency relationship: e.g. tar_target(downstream_target, f(upstream_target)) is a target named downstream_target which depends on a target upstream_target and a function f(). In addition, a target's name determines its random number generator seed. In this way, each target runs with a reproducible seed so someone else running the same pipeline should get the same results, and no two targets in the same pipeline share the same seed. (Even dynamic branches have different names and thus different seeds.) You can recover the seed of a completed target with tar_meta(your_target, seed) and run tar_seed_set() on the result to locally recreate the target's initial RNG state.

R code that runs in the format = "file" target and returns the file to be tracked.

R code to read the file. Must include !!.x where the file path goes: for example, read = readr::read_csv(file = !!.x, col_types = readr::cols()).

Logical, whether to enable tidy evaluation when interpreting command and pattern. If TRUE, you can use the "bang-bang" operator ⁠!!⁠ to programmatically insert the values of global objects.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Optional storage format for the target's return value. With the exception of format = "file", each target gets a file in ⁠_targets/objects⁠, and each format is a different way to save and load this file. See the "Storage formats" section for a detailed list of possible data storage formats.

Storage format of the file target, either "file" or "file_fast".

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Name of the target. tar_files() expects a unevaluated symbol for the name argument and an unevaluated expression for command, whereas tar_files_raw() expects a character string for the name argument and an evaluated expression object for command. See the examples for a demo.

R command for the target. tar_files() expects a unevaluated symbol for the name argument and an unevaluated expression for command, whereas tar_files_raw() expects a character string for the name argument and an evaluated expression object for command. See the examples for a demo.

Logical, whether to enable tidy evaluation when interpreting command and pattern. If TRUE, you can use the "bang-bang" operator ⁠!!⁠ to programmatically insert the values of global objects.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Character of length 1. Must be "file", "url", or "aws_file". See the format argument of targets::tar_target() for details.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, name of the iteration mode of the target. Choices:

• "vector": branching happens with vctrs::vec_slice() and aggregation happens with vctrs::vec_c().

• "list", branching happens with ⁠[[]]⁠ and aggregation happens with list().

• "group": dplyr::group_by()-like functionality to branch over subsets of a non-dynamic data frame. For iteration = "group", the target must not by dynamic (the pattern argument of tar_target() must be left NULL). The target's return value must be a data frame with a special tar_group column of consecutive integers from 1 through the number of groups. Each integer designates a group, and a branch is created for each collection of rows in a group. See the tar_group() function to see how you can create the special tar_group column with dplyr::group_by().

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date. Only applies to the downstream target. The upstream target always runs.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Name of the target. tar_files_input() expects a unevaluated symbol for the name argument, whereas tar_files_input_raw() expects a character string for name. See the examples for a demo.

Nonempty character vector of known existing input files to track for changes.

Positive integer of length 1, number of batches to partition the files. The default is one file per batch (maximum number of batches) which is simplest to handle but could cause a lot of overhead and consume a lot of computing resources. Consider reducing the number of batches below the number of files for heavy workloads.

Character, either "file", "file_fast", or "url". See the format argument of targets::tar_target() for details.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character, iteration method. Must be a method supported by the iteration argument of targets::tar_target(). The iteration method for the upstream target is always "list" in order to support batching.

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date. Only applies to the downstream target. The upstream target always runs.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Symbol, name of the target. In tar_target(), name is an unevaluated symbol, e.g. tar_target(name = data). In tar_target_raw(), name is a character string, e.g. tar_target_raw(name = "data").

A target name must be a valid name for a symbol in R, and it must not start with a dot. Subsequent targets can refer to this name symbolically to induce a dependency relationship: e.g. tar_target(downstream_target, f(upstream_target)) is a target named downstream_target which depends on a target upstream_target and a function f(). In addition, a target's name determines its random number generator seed. In this way, each target runs with a reproducible seed so someone else running the same pipeline should get the same results, and no two targets in the same pipeline share the same seed. (Even dynamic branches have different names and thus different seeds.) You can recover the seed of a completed target with tar_meta(your_target, seed) and run tar_seed_set() on the result to locally recreate the target's initial RNG state.

R code to run the target. In tar_target(), command is an unevaluated expression, e.g. tar_target(command = data). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(command = quote(data)).

R code for the condition that forces a build. If it evaluates to TRUE, then your work will run during tar_make().

Whether to invoke tidy evaluation (e.g. the ⁠!!⁠ operator from rlang) as soon as the target is defined (before tar_make()). Applies to arguments command and force.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Optional storage format for the target's return value. With the exception of format = "file", each target gets a file in ⁠_targets/objects⁠, and each format is a different way to save and load this file. See the "Storage formats" section for a detailed list of possible data storage formats.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, name of the iteration mode of the target. Choices:

• "vector": branching happens with vctrs::vec_slice() and aggregation happens with vctrs::vec_c().

• "list", branching happens with ⁠[[]]⁠ and aggregation happens with list().

• "group": dplyr::group_by()-like functionality to branch over subsets of a non-dynamic data frame. For iteration = "group", the target must not by dynamic (the pattern argument of tar_target() must be left NULL). The target's return value must be a data frame with a special tar_group column of consecutive integers from 1 through the number of groups. Each integer designates a group, and a branch is created for each collection of rows in a group. See the tar_group() function to see how you can create the special tar_group column with dplyr::group_by().

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date. Only applies to the downstream target. The upstream target always runs.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Character string, compression type for saving the data. See the compression argument of nanoparquet::write_parquet() for details.

Character vector with the data frame subclasses to assign. See the class argument of nanoparquet::parquet_options() for details.

Symbol, name of the target. In tar_target(), name is an unevaluated symbol, e.g. tar_target(name = data). In tar_target_raw(), name is a character string, e.g. tar_target_raw(name = "data").

A target name must be a valid name for a symbol in R, and it must not start with a dot. Subsequent targets can refer to this name symbolically to induce a dependency relationship: e.g. tar_target(downstream_target, f(upstream_target)) is a target named downstream_target which depends on a target upstream_target and a function f(). In addition, a target's name determines its random number generator seed. In this way, each target runs with a reproducible seed so someone else running the same pipeline should get the same results, and no two targets in the same pipeline share the same seed. (Even dynamic branches have different names and thus different seeds.) You can recover the seed of a completed target with tar_meta(your_target, seed) and run tar_seed_set() on the result to locally recreate the target's initial RNG state.

R code to run the target. In tar_target(), command is an unevaluated expression, e.g. tar_target(command = data). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(command = quote(data)).

Code to define a dynamic branching branching for a target. In tar_target(), pattern is an unevaluated expression, e.g. tar_target(pattern = map(data)). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(pattern = quote(map(data))).

To demonstrate dynamic branching patterns, suppose we have a pipeline with numeric vector targets x and y. Then, tar_target(z, x + y, pattern = map(x, y)) implicitly defines branches of z that each compute x[1] + y[1], x[2] + y[2], and so on. See the user manual for details.

Logical, whether to enable tidy evaluation when interpreting command and pattern. If TRUE, you can use the "bang-bang" operator ⁠!!⁠ to programmatically insert the values of global objects.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, name of the iteration mode of the target. Choices:

• "vector": branching happens with vctrs::vec_slice() and aggregation happens with vctrs::vec_c().

• "list", branching happens with ⁠[[]]⁠ and aggregation happens with list().

• "group": dplyr::group_by()-like functionality to branch over subsets of a non-dynamic data frame. For iteration = "group", the target must not by dynamic (the pattern argument of tar_target() must be left NULL). The target's return value must be a data frame with a special tar_group column of consecutive integers from 1 through the number of groups. Each integer designates a group, and a branch is created for each collection of rows in a group. See the tar_group() function to see how you can create the special tar_group column with dplyr::group_by().

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Character string, compression type for saving the data. See the compression argument of nanoparquet::write_parquet() for details.

Character vector with the data frame subclasses to assign. See the class argument of nanoparquet::parquet_options() for details.

Symbol, name of the target. In tar_target(), name is an unevaluated symbol, e.g. tar_target(name = data). In tar_target_raw(), name is a character string, e.g. tar_target_raw(name = "data").

A target name must be a valid name for a symbol in R, and it must not start with a dot. Subsequent targets can refer to this name symbolically to induce a dependency relationship: e.g. tar_target(downstream_target, f(upstream_target)) is a target named downstream_target which depends on a target upstream_target and a function f(). In addition, a target's name determines its random number generator seed. In this way, each target runs with a reproducible seed so someone else running the same pipeline should get the same results, and no two targets in the same pipeline share the same seed. (Even dynamic branches have different names and thus different seeds.) You can recover the seed of a completed target with tar_meta(your_target, seed) and run tar_seed_set() on the result to locally recreate the target's initial RNG state.

R code to run the target. In tar_target(), command is an unevaluated expression, e.g. tar_target(command = data). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(command = quote(data)).

Symbols, variables in the output data frame to group by.

Logical, whether to enable tidy evaluation when interpreting command and pattern. If TRUE, you can use the "bang-bang" operator ⁠!!⁠ to programmatically insert the values of global objects.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Optional storage format for the target's return value. With the exception of format = "file", each target gets a file in ⁠_targets/objects⁠, and each format is a different way to save and load this file. See the "Storage formats" section for a detailed list of possible data storage formats.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Symbol, name of the target. In tar_target(), name is an unevaluated symbol, e.g. tar_target(name = data). In tar_target_raw(), name is a character string, e.g. tar_target_raw(name = "data").

A target name must be a valid name for a symbol in R, and it must not start with a dot. Subsequent targets can refer to this name symbolically to induce a dependency relationship: e.g. tar_target(downstream_target, f(upstream_target)) is a target named downstream_target which depends on a target upstream_target and a function f(). In addition, a target's name determines its random number generator seed. In this way, each target runs with a reproducible seed so someone else running the same pipeline should get the same results, and no two targets in the same pipeline share the same seed. (Even dynamic branches have different names and thus different seeds.) You can recover the seed of a completed target with tar_meta(your_target, seed) and run tar_seed_set() on the result to locally recreate the target's initial RNG state.

R code to run the target. In tar_target(), command is an unevaluated expression, e.g. tar_target(command = data). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(command = quote(data)).

Positive integer, maximum number of row groups

Logical, whether to enable tidy evaluation when interpreting command and pattern. If TRUE, you can use the "bang-bang" operator ⁠!!⁠ to programmatically insert the values of global objects.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Optional storage format for the target's return value. With the exception of format = "file", each target gets a file in ⁠_targets/objects⁠, and each format is a different way to save and load this file. See the "Storage formats" section for a detailed list of possible data storage formats.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".

Symbol, name of the target. In tar_target(), name is an unevaluated symbol, e.g. tar_target(name = data). In tar_target_raw(), name is a character string, e.g. tar_target_raw(name = "data").

A target name must be a valid name for a symbol in R, and it must not start with a dot. Subsequent targets can refer to this name symbolically to induce a dependency relationship: e.g. tar_target(downstream_target, f(upstream_target)) is a target named downstream_target which depends on a target upstream_target and a function f(). In addition, a target's name determines its random number generator seed. In this way, each target runs with a reproducible seed so someone else running the same pipeline should get the same results, and no two targets in the same pipeline share the same seed. (Even dynamic branches have different names and thus different seeds.) You can recover the seed of a completed target with tar_meta(your_target, seed) and run tar_seed_set() on the result to locally recreate the target's initial RNG state.

R code to run the target. In tar_target(), command is an unevaluated expression, e.g. tar_target(command = data). In tar_target_raw(), command is an evaluated expression, e.g. tar_target_raw(command = quote(data)).

Tidyselect semantics to specify variables to group over. Alternatively, you can supply a character vector.

Logical, whether to enable tidy evaluation when interpreting command and pattern. If TRUE, you can use the "bang-bang" operator ⁠!!⁠ to programmatically insert the values of global objects.

Character vector of packages to load right before the target runs or the output data is reloaded for downstream targets. Use tar_option_set() to set packages globally for all subsequent targets you define.

Character vector of library paths to try when loading packages.

Optional storage format for the target's return value. With the exception of format = "file", each target gets a file in ⁠_targets/objects⁠, and each format is a different way to save and load this file. See the "Storage formats" section for a detailed list of possible data storage formats.

Character of length 1, remote repository for target storage. Choices:

• "local": file system of the local machine.

• "aws": Amazon Web Services (AWS) S3 bucket. Can be configured with a non-AWS S3 bucket using the endpoint argument of tar_resources_aws(), but versioning capabilities may be lost in doing so. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• "gcp": Google Cloud Platform storage bucket. See the cloud storage section of https://books.ropensci.org/targets/data.html for details for instructions.

• A character string from tar_repository_cas() for content-addressable storage.

Note: if repository is not "local" and format is "file" then the target should create a single output file. That output file is uploaded to the cloud and tracked for changes where it exists in the cloud. The local file is deleted after the target runs.

Character of length 1, what to do if the target stops and throws an error. Options:

• "stop": the whole pipeline stops and throws an error.

• "continue": the whole pipeline keeps going.

• "null": The errored target continues and returns NULL. The data hash is deliberately wrong so the target is not up to date for the next run of the pipeline.

• "abridge": any currently running targets keep running, but no new targets launch after that.

• "trim": all currently running targets stay running. A queued target is allowed to start if:

  1. It is not downstream of the error, and

  2. It is not a sibling branch from the same tar_target() call (if the error happened in a dynamic branch).

  The idea is to avoid starting any new work that the immediate error impacts. error = "trim" is just like error = "abridge", but it allows potentially healthy regions of the dependency graph to begin running. (Visit https://books.ropensci.org/targets/debugging.html to learn how to debug targets using saved workspaces.)

Character of length 1, memory strategy. If "persistent", the target stays in memory until the end of the pipeline (unless storage is "worker", in which case targets unloads the value from memory right after storing it in order to avoid sending copious data over a network). If "transient", the target gets unloaded after every new target completes. Either way, the target gets automatically loaded into memory whenever another target needs the value. For cloud-based dynamic files (e.g. format = "file" with repository = "aws"), this memory strategy applies to the temporary local copy of the file: "persistent" means it remains until the end of the pipeline and is then deleted, and "transient" means it gets deleted as soon as possible. The former conserves bandwidth, and the latter conserves local storage.

Logical, whether to run base::gc() just before the target runs.

Character of length 1. If deployment is "main", then the target will run on the central controlling R process. Otherwise, if deployment is "worker" and you set up the pipeline with distributed/parallel computing, then the target runs on a parallel worker. For more on distributed/parallel computing in targets, please visit https://books.ropensci.org/targets/crew.html.

Numeric of length 1 between 0 and 1. Controls which targets get deployed first when multiple competing targets are ready simultaneously. Targets with priorities closer to 1 get dispatched earlier (and polled earlier in tar_make_future()).

Object returned by tar_resources() with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities of targets. See tar_resources() for details.

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's return value is sent back to the host machine and saved/uploaded locally.

• "worker": the worker saves/uploads the value.

• "none": almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language. If you do use it, then the return value of the target is totally ignored when the target ends, but each downstream target still attempts to load the data file (except when retrieval = "none").

  If you select storage = "none", then the return value of the target's command is ignored, and the data is not saved automatically. As with dynamic files (format = "file") it is the responsibility of the user to write to the data store from inside the target.

  The distinguishing feature of storage = "none" (as opposed to format = "file") is that in the general case, downstream targets will automatically try to load the data from the data store as a dependency. As a corollary, storage = "none" is completely unnecessary if format is "file".

Character of length 1, only relevant to tar_make_clustermq() and tar_make_future(). Must be one of the following values:

• "main": the target's dependencies are loaded on the host machine and sent to the worker before the target runs.

• "worker": the worker loads the targets dependencies.

• "none": the dependencies are not loaded at all. This choice is almost never recommended. It is only for niche situations, e.g. the data needs to be loaded explicitly from another language.

An optional object from tar_cue() to customize the rules that decide whether the target is up to date.

Character of length 1, a custom free-form human-readable text description of the target. Descriptions appear as target labels in functions like tar_manifest() and tar_visnetwork(), and they let you select subsets of targets for the names argument of functions like tar_make(). For example, tar_manifest(names = tar_described_as(starts_with("survival model"))) lists all the targets whose descriptions start with the character string "survival model".