purrr (development version)
As of purrr 1.0.0, the
map()family of functions wraps all errors generated by
.finside an wrapper error that tracks the iteration index. As of purrr 1.0.1, this error now has a custom class (
namefields, and is documented in
map()errors with named inputs also report the name of the element that errored.
Fixed an issue where progress bars weren't being closed when user interrupts or errors were encountered during a
Fixed an invalid C signature for
Biarch: trueto build purrr on 32-bit Windows on R < 4.2.0 (#1017).
Core purpose refinements
cross()and all its variants have been deprecated in favour of
tidyr::expand_grid(). These functions were slow and buggy and we no longer think they are the right approach to solving this problem. See #768 for more information.
rerun()(#877), and the use of tidyselect with
map_at()and friends (#874) have been deprecated. These functions use some form of non-standard evaluation which we now believe is a poor fit for purrr.
lift_*family of functions has been deprecated. We no longer believe these to be a good fit for purrr because they rely on a style of function manipulation that is very uncommon in R code (#871).
list_along()have all been deprecated (#925). It's now clear that they don't align with the core purpose of purrr.
splice()is deprecated because we no longer believe that automatic splicing makes for good UI. Instead use
Use of map functions with expressions, calls, and pairlists has been deprecated (#961).
_raw()variants have been deprecated because they are of limited use and you can now use
map_chr(), automatic conversion from logical, integer, and double to character is now deprecated. Use an explicit
as.character()if needed (#904).
.fare now wrapped in an additional class that gives information about where the error occurred (#945).
Deprecation next steps
partial()are no longer supported. They have been defunct for quite some time.
Soft deprecated functions:
accumulate_right()are now fully deprecated. Similarly, the
partial(), and the
detect_index()are fully deprecated. Removing elements with
list_merge()is now fully deprecated.
is_scalar_numeric()have been removed. They have been deprecated since purrr 0.2.3 (Sep 2017).
invoke_*()is now deprecated. It was superseded in 0.3.0 (Jan 2019) and 3.5 years later, we have decided to deprecate it as part of the API refinement in the 1.0.0 release.
map_call()has been removed. It was made defunct in 0.3.0 (Jan 2019).
*_at()can now take a function (or formula) that's passed the vector of element names and returns the elements to select.
pmap_vec()work on all types of vectors, extending
map_int(), and friends so that you can easily work with dates, factors, date-times and more (#435).
discard_at()that work like
discard()but operation on element names rather than element contents (#817).
Some mapping functions have now a
.progressargument to create a progress bar. See
purrr is now licensed as MIT (#805).
modify2()are no longer generics. We have discovered a simple implementation that no longer requires genericity and methods were only provided by a very small number of packages (#894).
purrr now uses the base pipe (
|>) and anonymous function short hand (
\(x)), in all examples. This means that examples will no longer work in R 4.0 and earlier so in those versions of R, the examples are automatically converted to a regular section with a note that they might not work (#936).
When map functions fail, they now report the element they failed at (#945).
modify_tree()for recursively modifying nested data structures (#720).
Flattening and simplification
list_cbind()make it easy to
cbind()all of the elements in a list.
list_simplify()reduces a list of length-1 vectors to a simpler atomic or S3 vector (#900).
list_transpose()which automatically simplifies if possible (#875).
accumulate2()now both simplify the output if possible using vctrs. New arguments
ptypeallow you to control the details of simplification (#774, #809).
flatten()and friends are superseded in favour of
*_dfr()have been superseded in favour of using the appropriate map function along with
as_vector()have been superseded in favour of
list_simplify(). It provides a more consistent definition of simplification (#900).
transpose()has been superseded in favour of
list_transpose()(#875). It has built-in simplification.
_dbl()now use the same (strict) coercion methods as vctrs (#904). This means that:
map_chr(0L, identity), and
map_chr(1L, identity)are deprecated because we now believe that converting a logical/integer/double to a character vector should require an explicit coercion.
map_int(1.5, identity)now fails because we believe that silently truncating doubles to integers is dangerous. But note that
map_int(1, identity)still works since no numeric precision is lost.
map_int(c(TRUE, FALSE), identity),
map_dbl(c(TRUE, FALSE), identity),
map_lgl(c(1L, 0L), identity)and
map_lgl(c(1, 0), identity)now succeed because 1/TRUE and 0/FALSE should be interchangeable.
pmap()now use tidyverse recycling rules where vectors of length 1 are recycled to any size but all others must have the same length (#878).
pmap()now recycle names of their first input if needed (#783).
modify_at()have been reimplemented using vctrs principles. This shouldn't have an user facing impact, but it does make the implementation much simpler.
pluck_depth()and works with more types of input (#818).
pluck()now requires indices to be length 1 (#813). It also now reports the correct type if you supply an unexpected index.
pluck()now accepts negative integers, indexing from the right (#603).
chuck()now fail if you provide named inputs to ... (#788).
pluck()no longer replaces 0-length vectors with
default; it now only applies absent and
assign_in()can now modify non-existing locations (#704).
Setting with NULL
assign_in()now sets elements to
NULLrather than removing them (#636). Now use the explicit
zap()if you want to remove elements.
modify_if()now correctly handle
NULLs in replacement values (#655, #746, #753).
list_modify()'s interface has been standardised. Modifying with
NULLnow always creates a
NULLin the output (#810)
list_assign()which is similar to
list_modify()but doesn't work recursively (#822).
list_modify()no longer recurses into data frames (and other objects built on top of lists that are fundamentally non-list like) (#810). You can revert to the previous behaviour by setting
.is_node = is.list.
Minor improvements and bug fixes
conditionMessage()instead of directly interrogating the
modify()no longer works with calls or pairlists.
modify_depth()is no longer a generic. This makes it more consistent with
modify_depth()have a new
is_nodeargument that allows you to control what counts as a level. The default uses
vec_is_list()to avoid recursing into rich S3 objects like linear models or data.frames (#958, #920).
modify_depth()now correctly recurse at depth 1.
as_mapper()is now around twice as fast when used with character, integer, or list (#820).
modify_if(.else)is now actually evaluated for atomic vectors (@mgirlich, #701).
every()now correctly propagates missing values using the same rules as
&&(#751). Internally, it has become a wrapper around
&&. This makes it consistent with
&&and also with
some()which has always been a wrapper around
||with the same propagation rules.
some()now properly check the return value of their predicate function. It must now return a
Greatly improved performance of functions created with
partial()(#715). Their invocation is now as fast as for functions creating manually.
partial()no longer inlines the function in the call stack. This fixes issues when
partial()is used with
lm()for instance (#707).
- Fixes for CRAN checks.
Fixed issue in
list_modify()that prevented lists from being removed with
Added documentation for exporting functions created with purrr adverb (@njtierney, #668). See
none(), which tests that a predicate is false for all elements (the opposite of
every()) (@AliciaSchep, #735).
The documentation of
map()and its variants has been improved by @surdina as part of the Tidyverse Developer Day (@surdina, #671).
purrr now depends on R 3.2 or greater.
- Fix protection issues reported by rchk.
reduce()now forces arguments (#643).
Fixed an issue in
partial()with generic functions (#647).
negate()now works with generic functions and functions with early returns.
compose()now works with generic functions again (#629, #639). Its set of unit tests was expanded to cover many edge cases.
prepend()now works with empty lists (@czeildi, #637)
modify()and variants are now wrapping
[<-. This change increases the genericity of these functions but might cause different behaviour in some cases.
For instance, the
[[<-for data frames is stricter than the
[<-method and might throw errors instead of warnings. This is the case when assigning a longer vector than the number of rows.
[<-truncates the vector with a warning,
[[<-fails with an error (as is appropriate).
modify()and variants now return the same type as the input when the input is an atomic vector.
All functionals taking predicate functions (like
some()) got stricter. Predicate functions must now return a single
This change is meant to detect problems early with a more meaningful error message.
chuck()function. This is a strict variant of
pluck()that throws errors when an element does not exist instead of returning
pluck<-functions. They modify a data structure at an existing pluck location.
modify_in()function to map a function at a pluck location.
pluck()now dispatches properly with S3 vectors. The vector class must implement a
length()method for numeric indexing and a
names()method for string indexing.
pluck()now supports primitive functions (#404).
modify_if(). They take an alternative function that is mapped over elements of the input for which the predicate function returns
accumulate2()now terminate early when the function returns a value wrapped with
done()(#253). When an empty
done()is returned, the value at the last iteration is returned instead.
Functions taking predicates (
keep(), etc) now fail with an informative message when the return value is not
This is a breaking change for
some()which were documented to be more liberal in the values they accepted as logical (any vector was considered
TRUEif not a single
FALSEvalue, no matter its length). These functions signal soft-deprecation warnings instead of a hard failure.
Edit (purr 0.4.0):
some()never issued deprecation warnings because of a technical issue. We didn't fix the warnings in the end, and using predicates returning
NAis no longer considered deprecated. If you need to use
some()in contexts where
NApropagation is unsafe, e.g. in
if ()conditions, make sure to use safe predicate functions like
modify()and variants are now implemented using
[[<-methods. This implementation should be compatible with most vector classes.
imodify()functions. These work like
imap()but preserve the type of
.xin the return value.
pwalk()now preserve class for inputs of
POSIXctand other atomic S3 classes with an appropriate
[[method (#358, @mikmart).
modify_at()now preserve the class of atomic vectors instead of promoting them to lists. New S3 methods are provided for character, logical, double, and integer classes (@t-kalinowski, #417).
By popular request,
at_depth()has been brought back as
modify_depth(), it applies a function at a specified level of a data structure. However, it transforms all traversed vectors up to
.depthto bare lists (#381).
lmap_at()accept negative values for
.at, ignoring elements at those positions.
modify()now work with calls and pairlists (#412).
modify_depth()now modifies atomic leaves as well. This makes
modify_depth(x, 1, fn)equivalent to
accumulate2()function which is to
rate_delay()functions to create rate objects. You can pass rates to
slowly(), or the lower level function
rate_sleep(). This will cause a function to wait for a given amount of time with exponential backoff (increasingly larger waiting times) or for a constant delay.
insistently(f)modifies a function,
f, so that it is repeatedly called until it succeeds (@richierocks, @ijlyttle).
slowly()modifies a function so that it waits for a given amount of time between calls.
The interface of
partial() has been simplified. It now supports
quasiquotation to control the timing of evaluation, and the
rlang::call_modify() syntax to control the position of partialised
partial()now supports empty
... =argument to specify the position of future arguments, relative to partialised ones. This syntax is borrowed from (and implemented with)
To prevent partial matching of
...f, the latter has been renamed to
.f, which is more consistent with other purrr function signatures.
partial()now supports quasiquotation. When you unquote an argument, it is evaluated only once at function creation time. This is more flexible than the
.lazyargument since you can control the timing of evaluation for each argument. Consequently,
.lazyis soft-deprecated (#457).
Fixed an infinite loop when partialised function is given the same name as the original function (#387).
as_closure()on primitive functions to ensure argument matching (#360).
partial()is soft-deprecated in favour of quasiquotation:
# Before partial(fn, u = runif(1), n = rnorm(1), .lazy = FALSE) # After partial(fn, u = !!runif(1), n = !!rnorm(1)) # All constant partial(fn, u = !!runif(1), n = rnorm(1)) # First constant
Minor improvements and fixes
The tibble package is now in Suggests rather than Imports. This brings the hard dependency of purrr to just rlang and magrittr.
compose()now returns an identity function when called without inputs.
Functions created with
compose()now have the same formal parameters as the first function to be called. They also feature a more informative print method that prints all composed functions in turn (@egnha, #366).
compose(). When set to
"forward", the functions are composed from left to right rather than right to left.
list_modify()now supports the
zap()sentinel (reexported from rlang) to remove elements from lists. Consequently, removing elements with the ambiguous sentinel
The requirements of
list_merge()have been relaxed. Previously it required both the modified lists and the inputs to be either named or unnamed. This restriction now only applies to inputs in
.... When inputs are all named, they are matched to the list by name. When they are all unnamed, they are matched positionally. Otherwise, this is an error.
Fixed ordering of names returned by
accumulate_right()output. They now correspond to the order of inputs.
Fixed names of
compose()now supports composition with lambdas (@ColinFay, #556)
pmap()crash with empty lists on the Win32 platform (#565).
.raggedargument evaluates correctly to
TRUEby default when
.depth < 0(@cderv, #530).
accumulate()now inherits names from their first input (@AshesITR, #446).
attr_getter()no longer uses partial matching. For example, if an
xobject has a
labelsattribute but no
attr_getter("label")(x)will no longer extract the
labelsattribute (#460, @huftis).
flatten_dfc()now aborts if dplyr is not installed. (#454)
imap_dfr()now works with
.idargument is provided (#429)
list_merge()now handle duplicate duplicate argument names correctly (#441, @mgirlich).
pmap_rawadded to support raw vectors. (#455, @romainfrancois)
flatten()now supports raw and complex elements.
array_tree()now retain the
dimnames()of the input array (#584, @flying-sheep)
pluck()no longer flattens lists of arguments. You can still do it manually with
!!!. This change is for consistency with other dots-collecting functions of the tidyverse.
modify_at()now supports selection using
Note that for now you need to import
vars()from dplyr or call it qualified like
dplyr::vars(). It will be reexported from rlang in a future release.
detect()now has a .default argument to specify the value returned when nothing is detected (#622, @ColinFay).
We have standardised the purrr API for reverse iteration with a common
reduce_right()is soft-deprecated and replaced by a new
# Before: reduce_right(1:3, f) # After: reduce(1:3, f, .dir = "backward")
Note that the details of the computation have changed. Whereas
f(f(3, 2), 1), it now computes
f(1, f(2, 3)). This is the standard way of reducing from the right.
To produce the exact same reduction as
reduce_right(), simply reverse your vector and use a left reduction:
# Before: reduce_right(1:3, f) # After: reduce(rev(1:3), f)
reduce2_right()is soft-deprecated without replacement. It is not clear what algorithmic properties should a right reduction have in this case. Please reach out if you know about a use case for a right reduction with a ternary function.
accumulate_right()is soft-deprecated and replaced by the new
accumulate(). Note that the algorithm has slightly changed: the accumulated value is passed to the right rather than the left, which is consistent with a right reduction.
# Before: accumulate_right(1:3, f) # After: accumulate(1:3, f, .dir = "backward")
detect_index()is soft-deprecated and renamed to
.dirfor consistency with other functions and clarity of the interface.
# Before detect(x, f, .right = TRUE) # After detect(x, f, .dir = "backward")
The interface of
partial() has been simplified (see more about
partial()is soft-deprecated in favour of quasiquotation.
We had to rename
partial()in order to support
... =argument (which would otherwise partial-match on
...f). This also makes
partial()more consistent with other purrr function signatures.
invoke_map() are retired in favour of
that retired functions are no longer under active development, but
continue to be maintained undefinitely in the package.
invoke()is retired in favour of the
exec()function, reexported from rlang.
exec()evaluates a function call built from its inputs and supports tidy dots:
# Before: invoke(mean, list(na.rm = TRUE), x = 1:10) # After exec(mean, 1:10, !!!list(na.rm = TRUE))
Note that retired functions are not removed from the package and will be maintained undefinitely.
invoke_map()is retired without replacement because it is more complex to understand than the corresponding code using
# Before: invoke_map(fns, list(args)) invoke_map(fns, list(args1, args2)) # After: map(fns, exec, !!!args) map2(fns, list(args1, args2), function(fn, args) exec(fn, !!!args))
Other lifecycle changes
%@%is soft-deprecated, please use the operator exported in rlang instead. The latter features an interface more consistent with
@as it uses NSE, supports S4 fields, and has an assignment variant.
Removing elements from lists using
list_modify()is soft-deprecated. Please use the new
zap()sentinel reexported from rlang instead:
# Before: list_modify(x, foo = NULL) # After: list_modify(x, foo = zap())
This change is motivated by the ambiguity of
NULLas a deletion sentinel because
NULLis also a valid value in lists. In the future,
NULLwill set an element to
NULLrather than removing the element.
rerun()is now in the questioning stage because we are no longer convinced NSE functions are a good fit for purrr. Also,
rerun(n, x)can just as easily be expressed as
map(1:n, ~ x)(with the added benefit of being passed the current index as argument to the lambda).
- This is a maintenance release following the release of dplyr 0.7.5.
- Fixes for R 3.1.
We noticed the following issues during reverse dependencies checks:
reduce()fails with this message:
Error: `.x` is empty, and no `.init` supplied, this is because
.xis empty. Fix the problem by supplying an appropriate argument to
.init, or by providing special behaviour when
.xhas length 0.
The type predicates have been migrated to rlang. Consequently the
bare-type-predicatesdocumentation topic is no longer in purrr, which might cause a warning if you cross-reference it.
purrr no longer depends on lazyeval or Rcpp (or dplyr, as of the previous version). This makes the dependency graph of the tidyverse simpler, and makes purrr more suitable as a dependency of lower-level packages.
There have also been two changes to eliminate name conflicts between purrr and dplyr:
split_by()have been removed.
dplyr::order_by()and the complete family doesn't feel that useful. Use tibbles instead (#217).
contains()has been renamed to
has_element()to avoid conflicts with dplyr (#217).
The plucking mechanism used for indexing into data structures with
map() has been extracted into the function
pluck(). Plucking is
often more readable to extract an element buried in a deep data
structure. Compare this syntax-heavy extraction which reads
to the equivalent pluck:
x %>% pluck(1, accessor, "foo")
as_mapper()because it is a tranformation that makes sense primarily for mapping functions, not in general (#298).
.nullhas been renamed to
.defaultto better reflect its intent (#298).
.defaultis returned whenever an element is absent or empty (#231, #254).
as_mapper()sanitises primitive functions by transforming them to closures with standardised argument names (using
rlang::as_closure()). For instance
+is transformed to
function(.x, .y) .x + .y. This results in proper argument matching so that
, .x = 5))produces
list(5 - 1, 5 - 2, ...).
Recursive indexing can now extract objects out of environments (#213) and S4 objects (#200), as well as lists.
attr_getter()makes it possible to extract from attributes like
map(list(iris, mtcars), attr_getter("row.names")).
The argument list for formula-functions has been tweaked so that you can refer to arguments by position with
..2, and so on. This makes it possible to use the formula shorthand for functions with more than two arguments (#289).
safely()and friends no longer capture interrupts: this means that you can now terminate a mapper using one of these with Escape or Ctrl + C (#314)
All map functions now treat
NULLthe same way as an empty vector (#199), and return an empty vector if any input is an empty vector.
map()functions now force their arguments in the same way that base R does for
lapply()(#191). This makes
map()etc easier to use when generating functions.
A new family of "indexed" map functions,
imap_lgl()etc, provide a short-hand for
The data frame suffix
_dfhas been (soft) deprecated in favour of
_dfrto more clearly indicate that it's a row-bind. All variants now also have a
_dfcfor column binding (#167). (These will not be terribly useful until
dplyr::bind_cols()have better semantics for vectors.)
modify() family returns the same output of the type as the
.x. This is in contrast to the
map() family which always
returns a list, regardless of the input type.
The modify functions are S3 generics. However their default methods
should be sufficient for most classes since they rely on the semantics
modify.default() is thus a shorthand for
x <- map(x, f).
at_depth()has been renamed to
.raggedargument, and negative depths are now computed relative to the deepest component of the list (#236).
auto_browse(f)returns a new function that automatically calls
fthrows an error (#281).
vec_depth()computes the depth (i.e. the number of levels of indexing) or a vector (#243).
reduce2_right()make it possible to reduce with a 3 argument function where the first argument is the accumulated value, the second argument is
.x, and the third argument is
stats::modifyList()to replace by position if the list is not named.(#201).
list_merge()operates similarly to
list_modify()but combines instead of replacing (#322).
The legacy function
update_list()is basically a version of
list_modifythat evaluates formulas within the list. It is likely to be deprecated in the future in favour of a tidyeval interface such as a list method for
Minor improvements and bug fixes
Thanks to @dchiu911, the unit test coverage of purrr is now much greater.
All predicate functions are re-exported from rlang (#124).
compact()now works with standard mapper conventions (#282).
cross_n()has been renamed to
_nsuffix was removed for consistency with
map_n()at the start of the project) and
cross_d()has been renamed to
cross_df()for consistency with
NAif present in the input (#174).
invoke()uses a more robust approach to generate the argument list (#249) It no longer uses lazyeval to figure out which enviroment a character
is_scalar_numeric()are deprecated because they don't test for what you might expect at first sight.
reduce()now throws an error if
.xis empty and
.initis not supplied.
zip_n()have been removed.
pmap()coerces data frames to lists to avoid the expensive
[.data.framewhich provides security that is unneeded here (#220).
rdunif()checks its inputs for validity (#211).
set_names()can now take a function to tranform the names programmatically (#276), and you can supply names in
...to reduce typing even more more (#316).
set_names()is now powered by
safely()now actually uses the
transpose()now matches by name if available (#164). You can override the default choice with the new
The function argument of
detect_index()have been renamed from
.f. This is because they have mapper semantics rather than predicate semantics.
This is a compatibility release with dplyr 0.6.0.
- All data-frame based mappers have been removed in favour of new
functions and idioms in the tidyverse.
unslice()have been moved to purrrlyr. This is a bit of an aggresive change but it allows us to make the dependencies much lighter.
Fix for dev tibble support.
as_function()now supports list arguments which allow recursive indexing using either names or positions. They now always stop when encountering the first NULL (#173).
reducecorrectly pass extra arguments to the worker function.
.nullargument that for character and numeric values allows you to specify what to return for null/absent elements (#110). This can be used with any map function, e.g.
map_int(x, 1, .null = NA)
as_function()is now generic.
TRUEonly for regular functions.
Fix crash on GCC triggered by
There are two handy infix functions:
x %||% yis shorthand for
if (is.null(x)) y else x(#109).
x %@% "a"is shorthand for
attr(x, "a", exact = TRUE)(#69).
accumulate()has been added to handle recursive folding. It is shortand for
Reduce(f, .x, accumulate = TRUE)and follows a similar syntax to
reduce()(#145). A right-hand version
accumulate_right()was also added.
map_df()row-binds output together. It's the equivalent of
flatten()is now type-stable and always returns a list. To return a simpler vector, use
invoke()has been overhauled to be more useful: it now works similarly to
.xis NULL, and hence
map_call()has been deprecated.
invoke_map()is a vectorised complement to
invoke()(#125), and comes with typed variants
zip_n()(#128). The name more clearly reflects the intent (transposing the first and second levels of list). It no longer has fields argument or the
.simplifyargument; instead use the new
possibly()are experimental functions for working with functions with side-effects (e.g. printed output, messages, warnings, and errors) (#120).
safely()is a version of
try()that modifies a function (rather than an expression), and always returns a list with two components,
rep_along()generalise the idea of
is_null()is the snake-case version of
pmap()(parallel map) replaces
map_n()(#132), and has typed-variants suffixed
set_names()is a snake-case alternative to
setNames()with stricter equality checking, and more convenient defaults for pipes:
x %>% set_names()is equivalent to
Row based functionals
We are still figuring out what belongs in dplyr and what belongs in purrr. Expect much experimentation and many changes with these functions.
map()now always returns a list. Data frame support has been moved to
dmap(). The latter supports sliced data frames as a shortcut for the combination of
x %>% by_slice(dmap, fun, .collate = "rows"). The conditional variants
dmap_if()also support sliced data frames and will recycle scalar results to the slice size.
map_rows()has been renamed to
invoke_rows(). As other rows-based functionals, it collates results inside lists by default, but with column collation this function is equivalent to
The rows-based functionals gain a
.tooption to name the output column as well as a
.collateargument. The latter allows to collate the output in lists (by default), on columns or on rows. This makes these functions more flexible and more predictable.
Bug fixes and minor changes
as_function(), which converts formulas etc to functions, is now exported (#123).
rerun()is correctly scoped (#95)
update_list()can now modify an element called
map*()now use custom C code, rather than relying on
mapply()etc. The performance characteristcs are very similar, but it allows us greater control over the output (#118).
map_lgl()now has second argument
map()followed by the appropriate
map3(x, y, z)->
map_n(list(x, y, z));
walk3(x, y, z) ->pwalk(list(x, y, z))`