Flexibly restructure and aggregate data using just two functions: melt and 'dcast' (or 'acast').
reshape2 is retired: only changes necessary to keep it on CRAN will be made. We recommend using tidyr instead.
Reshape2 is a reboot of the reshape package. It's been over five years since the first release of reshape, and in that time I've learned a tremendous amount about R programming, and how to work with data in R. Reshape2 uses that knowledge to make a new package for reshaping data that is much more focused and much much faster.
This version improves speed at the cost of functionality, so I have renamed it to
reshape2 to avoid causing problems for existing users. Based on user feedback I may reintroduce some of these features.
What's new in
considerably faster and more memory efficient thanks to a much better underlying algorithm that uses the power and speed of subsetting to the fullest extent, in most cases only making a single copy of the data.
cast is replaced by two functions depending on the output type:
produces data frames, and
acast produces matrices/arrays.
multidimensional margins are now possible:
been dropped: now the name of the margin refers to the variable that has
its value set to (all).
some features have been removed such as the
| cast operator, and the
ability to return multiple values from an aggregation function. I'm
reasonably sure both these operations are better performed by plyr.
a new cast syntax which allows you to reshape based on functions of variables (based on the same underlying syntax as plyr):
better development practices like namespaces and tests.
melt now names the columns of its returned data frame
VarN instead of
melt replaces the old argument
Initial benchmarking has shown
melt to be up to 10x faster, pure reshaping
cast up to 100x faster, and aggregating
cast() up to 10x faster.
This work has been generously supported by BD (Becton Dickinson).
Fix C/C++ problems causing R CMD CHECK errors.
melt.data.frame() throws when encountering objects of type
and requests a conversion to the (much saner)
melt.data.frame() now properly sets the OBJECT bit on
generated if attributes are copied (for example, when multiple POSIXct
columns are concatenated to generate the
value variable) (#50)
melt.data.frame() can melt
list elements as
melt.data.frame() no longer errors when
NULL or empty.
acast() gain a useful error message if you use
value.var (#16), and if
value.var doesn't exist (#9). They
also work better with
. in specifications like
. ~ . or
x + y ~ .
melt.array() creates factor variables with levels in the same order
as the original rownames (#19)
melt.data.frame() gains an internal Rcpp / C++ implementation, and
is now many orders of magnitudes faster. It also preserves identical
attributes for measure variables, and now throws a warning if they are
dropped. (Thanks to Kevin Ushey)
melt.data.frame() gains a
factorsAsStrings argument that controls whether
factors are converted to character when melted as measure variables. This
TRUE by default for backward compatibility.
melt.array() gains a
as.is argument which can be used to prevent
dimnames being converted with
recast() now returns a data frame instead of a list (#45).
Fix incompatibility with plyr 1.8
Fix evaluation bug revealed by knitr. (Fixes #18)
Fixed a bug in
melt where it didn't automatically get variable names
when used with tables. (Thanks to Winston Chang)
Fixed bug in melt where factors were converted to integers, instead of to characters
When the measured variable is a factor,
dcast now converts it to a
character rather than throwing an error.
acast still returns a factor
matrix. (Thanks to Brian Diggs.)
acast is now much faster, due to fixing a very slow way of naming the
output. (Thanks to José Bartolomei Díaz for the bug report)
value_var argument to
dcast renamed to
value.var to be
consistent with other argument names
NA factor levels before
(all) when creating margins
Corrected reshape citation.
melt.data.frame no longer turns characters into factors
All melt methods gain a
value.name arguments - these
previously were only possessed by
melt.data.frame (Fixes #5)