Example of why the R function apply() is hard to grok and how to use plyr instead

New R users tend to struggle with the apply() family of functions. This video illustrates why the struggle and how the plyr package can help.

Category:

Education

Tags:

• rstats
• apply
• sapply
• plry

[contingocoins]

when you define function(x) as { GetValues (x) }

why not just call getValues?

i.e.

sapply(yearList, getValues)

[joeionnojitsu]

apply is for arrays, e.g., if you have a matrix and want to apply a function to each row or each column. The result is another array---*if* the function you used returned the same number of elements for each row/column/whatever. Otherwise it's a list (and a mess).

Bottom line: lapply is the safest one of the bunch. List (or vector) in, list out.

[joeionnojitsu]

sapply should always be used with care---it's basically lapply, followed by an attempt to simply the result into an array. That simplification process is what tends to go awry.

lapply is for lists---apply a function to each element of a list. The result is always a list.

tapply is basically split + lapply. You use it when you want a function to act on subsets of the input vector that are defined by a factor.