Sum of rows based on column value

Question

I want to sum rows that have the same value in one column:

> df <- data.frame("1"=c("a","b","a","c","c"), "2"=c(1,5,3,6,2), "3"=c(3,3,4,5,2))
> df
  X1 X2 X3
1  a  1  3
2  b  5  3
3  a  3  4
4  c  6  5
5  c  2  2

For one column (X2), the data can be aggregated to get the sums of all rows that have the same X1 value:

> ddply(df, .(X1), summarise, X2=sum(X2))
  X1 X2
1  a  4
2  b  5
3  c  8

How do I do the same for X3 and an arbitrary number of other columns except X1?

This is the result I want:

  X1 X2 X3
1  a  4  7
2  b  5  3
3  c  8  7

Answer 1

ddply(df, "X1", numcolwise(sum))

see ?numcolwise for details and examples.

Answer 2

aggregate can easily do this with the formula interface:

aggregate(. ~ X1, data=df, FUN=sum)
##   X1 X2 X3
## 1  a  4  7
## 2  b  5  3
## 3  c  8  7

Equivalently:

aggregate(cbind(X2, X3) ~ X1, data=df, FUN=sum)

Answer 3

aggregate is a great function for these sorts of things:

aggregate(df[,-1],df["X1"],sum)

  X1 X2 X3
1  a  4  7
2  b  5  3
3  c  8  7

And a base R version of the numcolwise method from plyr:

aggregate(df[,sapply(df,is.numeric)],df["X1"],sum)

Answer 4

A data.table solution for memory efficiency and coding elegance

library(data.table)
DT <- data.table(df)


DT[, lapply(.SD, sum), by = X1]

.SD is the subset of the data.table for each group defined by the values of X1. There are 3 helpful vignettes associated with the data.table package.