A target is a single step of computation in a pipeline. It runs an R command and returns a value. This value gets treated as an R object that can be used by the commands of targets downstream. Targets that are already up to date are skipped. See the user manual for more details.
tar_target()
defines a target using non-standard evaluation.
The name
argument is an unevaluated symbol,
and the command
and pattern
arguments are unevaluated expressions. Example:
tar_target(name = data, command = get_data())
.
tar_target_raw()
defines a target with standard evaluation.
The name
argument is a character string,
and the command
and pattern
arguments are evaluated expressions. Example:
tar_target_raw(name = "data", command = quote(get_data()))
.
tar_target_raw()
also has extra arguments deps
and string
for advanced customization.
Usage
tar_target(
name,
command,
pattern = NULL,
tidy_eval = targets::tar_option_get("tidy_eval"),
packages = targets::tar_option_get("packages"),
library = targets::tar_option_get("library"),
format = targets::tar_option_get("format"),
repository = targets::tar_option_get("repository"),
iteration = targets::tar_option_get("iteration"),
error = targets::tar_option_get("error"),
memory = targets::tar_option_get("memory"),
garbage_collection = targets::tar_option_get("garbage_collection"),
deployment = targets::tar_option_get("deployment"),
priority = targets::tar_option_get("priority"),
resources = targets::tar_option_get("resources"),
storage = targets::tar_option_get("storage"),
retrieval = targets::tar_option_get("retrieval"),
cue = targets::tar_option_get("cue"),
description = targets::tar_option_get("description")
)
tar_target_raw(
name,
command,
pattern = NULL,
packages = targets::tar_option_get("packages"),
library = targets::tar_option_get("library"),
deps = NULL,
string = NULL,
format = targets::tar_option_get("format"),
repository = targets::tar_option_get("repository"),
iteration = targets::tar_option_get("iteration"),
error = targets::tar_option_get("error"),
memory = targets::tar_option_get("memory"),
garbage_collection = targets::tar_option_get("garbage_collection"),
deployment = targets::tar_option_get("deployment"),
priority = targets::tar_option_get("priority"),
resources = targets::tar_option_get("resources"),
storage = targets::tar_option_get("storage"),
retrieval = targets::tar_option_get("retrieval"),
cue = targets::tar_option_get("cue"),
description = targets::tar_option_get("description")
)
Arguments
- name
Symbol, name of the target. In
tar_target()
,name
is an unevaluated symbol, e.g.tar_target(name = data)
. Intar_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 nameddownstream_target
which depends on a targetupstream_target
and a functionf()
. 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 withtar_meta(your_target, seed)
and runtar_seed_set()
on the result to locally recreate the target's initial RNG state.- command
R code to run the target. In
tar_target()
,command
is an unevaluated expression, e.g.tar_target(command = data)
. Intar_target_raw()
,command
is an evaluated expression, e.g.tar_target_raw(command = quote(data))
.- pattern
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))
. Intar_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
andy
. Then,tar_target(z, x + y, pattern = map(x, y))
implicitly defines branches ofz
that each computex[1] + y[1]
,x[2] + y[2]
, and so on. See the user manual for details.- tidy_eval
Logical, whether to enable tidy evaluation when interpreting
command
andpattern
. IfTRUE
, you can use the "bang-bang" operator!!
to programmatically insert the values of global objects.- packages
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.- library
Character vector of library paths to try when loading
packages
.- format
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.- repository
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 theendpoint
argument oftar_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"
andformat
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.- iteration
Character of length 1, name of the iteration mode of the target. Choices:
"vector"
: branching happens withvctrs::vec_slice()
and aggregation happens withvctrs::vec_c()
."list"
, branching happens with[[]]
and aggregation happens withlist()
."group"
:dplyr::group_by()
-like functionality to branch over subsets of a non-dynamic data frame. Foriteration = "group"
, the target must not by dynamic (thepattern
argument oftar_target()
must be leftNULL
). The target's return value must be a data frame with a specialtar_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 thetar_group()
function to see how you can create the specialtar_group
column withdplyr::group_by()
.
- error
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 returnsNULL
. 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:It is not downstream of the error, and
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 likeerror = "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.)
- memory
Character of length 1, memory strategy. If
"persistent"
, the target stays in memory until the end of the pipeline (unlessstorage
is"worker"
, in which casetargets
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"
withrepository = "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.- garbage_collection
Logical, whether to run
base::gc()
just before the target runs.- deployment
Character of length 1. If
deployment
is"main"
, then the target will run on the central controlling R process. Otherwise, ifdeployment
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 intargets
, please visit https://books.ropensci.org/targets/crew.html.- priority
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()
).- resources
Object returned by
tar_resources()
with optional settings for high-performance computing functionality, alternative data storage formats, and other optional capabilities oftargets
. Seetar_resources()
for details.- storage
Character of length 1, only relevant to
tar_make_clustermq()
andtar_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 whenretrieval = "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 toformat = "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 ifformat
is"file"
.
- retrieval
Character of length 1, only relevant to
tar_make_clustermq()
andtar_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.
- cue
An optional object from
tar_cue()
to customize the rules that decide whether the target is up to date.- description
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()
andtar_visnetwork()
, and they let you select subsets of targets for thenames
argument of functions liketar_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"
.- deps
Optional character vector of the adjacent upstream dependencies of the target, including targets and global objects. If
NULL
, dependencies are resolved automatically as usual. Thedeps
argument is only for developers of extension packages such astarchetypes
, not for end users, and it should almost never be used at all. In scenarios that at first appear to requiresdeps
, there is almost always a simpler and more robust workaround that avoids settingdeps
.- string
Optional string representation of the command. Internally, the string gets hashed to check if the command changed since last run, which helps
targets
decide whether the target is up to date. External interfaces can take control ofstring
to ignore changes in certain parts of the command. IfNULL
, the strings is just deparsed fromcommand
(default).
Value
A target object. Users should not modify these directly,
just feed them to list()
in your target script file
(default: _targets.R
).
Target objects
Functions like tar_target()
produce target objects,
special objects with specialized sets of S3 classes.
Target objects represent skippable steps of the analysis pipeline
as described at https://books.ropensci.org/targets/.
Please read the walkthrough at
https://books.ropensci.org/targets/walkthrough.html
to understand the role of target objects in analysis pipelines.
For developers, https://wlandau.github.io/targetopia/contributing.html#target-factories explains target factories (functions like this one which generate targets) and the design specification at https://books.ropensci.org/targets-design/ details the structure and composition of target objects.
Storage formats
targets
has several built-in storage formats to control how return
values are saved and loaded from disk:
"rds"
: Default, usessaveRDS()
andreadRDS()
. Should work for most objects, but slow."auto"
: either"file"
or"qs"
, depending on the return value of the target. If the return value is a character vector of existing files (and/or directories), then the format becomes"file"
beforetar_make()
saves the target. Otherwise, the format becomes"qs"
."qs"
: Usesqs::qsave()
andqs::qread()
. Should work for most objects, much faster than"rds"
. Optionally set the preset forqsave()
throughtar_resources()
andtar_resources_qs()
."feather"
: Usesarrow::write_feather()
andarrow::read_feather()
(version 2.0). Much faster than"rds"
, but the value must be a data frame. Optionally setcompression
andcompression_level
inarrow::write_feather()
throughtar_resources()
andtar_resources_feather()
. Requires thearrow
package (not installed by default)."parquet"
: Usesarrow::write_parquet()
andarrow::read_parquet()
(version 2.0). Much faster than"rds"
, but the value must be a data frame. Optionally setcompression
andcompression_level
inarrow::write_parquet()
throughtar_resources()
andtar_resources_parquet()
. Requires thearrow
package (not installed by default)."fst"
: Usesfst::write_fst()
andfst::read_fst()
. Much faster than"rds"
, but the value must be a data frame. Optionally set the compression level forfst::write_fst()
throughtar_resources()
andtar_resources_fst()
. Requires thefst
package (not installed by default)."fst_dt"
: Same as"fst"
, but the value is adata.table
. Deep copies are made as appropriate in order to protect against the global effects of in-place modification. Optionally set the compression level the same way as for"fst"
."fst_tbl"
: Same as"fst"
, but the value is atibble
. Optionally set the compression level the same way as for"fst"
."keras"
: superseded bytar_format()
and incompatible witherror = "null"
(intar_target()
ortar_option_set()
). Useskeras::save_model_hdf5()
andkeras::load_model_hdf5()
. The value must be a Keras model. Requires thekeras
package (not installed by default)."torch"
: superseded bytar_format()
and incompatible witherror = "null"
(intar_target()
ortar_option_set()
). Usestorch::torch_save()
andtorch::torch_load()
. The value must be an object from thetorch
package such as a tensor or neural network module. Requires thetorch
package (not installed by default)."file"
: A dynamic file. To use this format, the target needs to manually identify or save some data and return a character vector of paths to the data (must be a single file path ifrepository
is not"local"
). (These paths must be existing files and nonempty directories.) Then,targets
automatically checks those files and cues the appropriate run/skip decisions if those files are out of date. Those paths must point to files or directories, and they must not contain characters|
or*
. All the files and directories you return must actually exist, or elsetargets
will throw an error. (And ifstorage
is"worker"
,targets
will first stall out trying to wait for the file to arrive over a network file system.) If the target does not create any files, the return value should becharacter(0)
.If
repository
is not"local"
andformat
is"file"
, then the character vector returned by the target must be of length 1 and point to a single file. (Directories and vectors of multiple file paths are not supported for dynamic files on the cloud.) 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."url"
: A dynamic input URL. For this storage format,repository
is implicitly"local"
, URL format is likeformat = "file"
except the return value of the target is a URL that already exists and serves as input data for downstream targets. Optionally supply a customcurl
handle throughtar_resources()
andtar_resources_url()
. innew_handle()
,nobody = TRUE
is important because it ensurestargets
just downloads the metadata instead of the entire data file when it checks time stamps and hashes. The data file at the URL needs to have an ETag or a Last-Modified time stamp, or else the target will throw an error because it cannot track the data. Also, use extreme caution when trying to useformat = "url"
to track uploads. You must be absolutely certain the ETag and Last-Modified time stamp are fully updated and available by the time the target's command finishes running.targets
makes no attempt to wait for the web server.A custom format can be supplied with
tar_format()
. For this choice, it is the user's responsibility to provide methods for (un)serialization and (un)marshaling the return value of the target.The formats starting with
"aws_"
are deprecated as of 2022-03-13 (targets
version > 0.10.0). For cloud storage integration, use therepository
argument instead.
Formats "rds"
, "file"
, and "url"
are general-purpose formats
that belong in the targets
package itself.
Going forward, any additional formats should be implemented with
tar_format()
in third-party packages like tarchetypes
and geotargets
(for example: tarchetypes::tar_format_nanoparquet()
).
Formats "qs"
, "fst"
, etc. are legacy formats from before the
existence of tar_format()
, and they will continue to remain in
targets
without deprecation.
See also
Other targets:
tar_cue()
Examples
# Defining targets does not run them.
data <- tar_target(target_name, get_data(), packages = "tidyverse")
analysis <- tar_target(analysis, analyze(x), pattern = map(x))
# In a pipeline:
if (identical(Sys.getenv("TAR_EXAMPLES"), "true")) { # for CRAN
tar_dir({ # tar_dir() runs code from a temp dir for CRAN.
tar_script({
library(targets)
library(tarchetypes)
list(
tar_target(name = x, command = 1 + 1),
tar_target_raw(name = "y", command = quote(x + y))
)
})
tar_make()
tar_read(x)
})
# Tidy evaluation
tar_option_set(envir = environment())
n_rows <- 30L
data <- tar_target(target_name, get_data(!!n_rows))
print(data)
# Disable tidy evaluation:
data <- tar_target(target_name, get_data(!!n_rows), tidy_eval = FALSE)
print(data)
tar_option_reset()
}