collapse: Collapse a data.table by Group

In kevinushey/data.table.extras: Add-on functions to augment the data.table package.

Description

For each data.table subset (as subset by the by argument), return the rows at indices i.

Usage

collapse(DT, i = 1, by)

Arguments

DT	A data.table.
i	Indices at which to subset each sub-data.table.
by	Variable(s) to subset over, as a list of symbols.

Details

The code wraps to the impossible to remember but incredibly powerful and fast DT[ DT[, .I[i], by=by]$V1 ].

Examples

DT <- data.table(
  x=rep(letters, each=1E5),
  y=rnorm(26*1E5),
  z=rep(LETTERS, each=1E5),
  a=rep(LETTERS, each=1E5),
  b=rep(LETTERS, each=1E5),
  c=rep(LETTERS, each=1E5),
  d=rep(LETTERS, each=1E5),
  e=rep(LETTERS, each=1E5)
)

DT[, mean_y := mean(y), by=x]
collapse(DT, 1, by=list(x))

kevinushey/data.table.extras documentation built on May 20, 2019, 9:09 a.m.