Question

I have a list of sublists. Each sublist contains an identical data frame (identical except for the data inside it) and a ‘yes/no’ label. I’d like to find the row-wise mean of the data frames, if the yes/no label is TRUE.

#Create the data frames
id <- c("a", "b", "c")
df1 <- data.frame(id=id, data=c(1, 2, 3))
df2 <- df1
df3 <- data.frame(id=id, data=c(1000, 2000, 3000))

#Create the sublists that will store the data frame and the yes/no variable
sub1 <- list(data=df1, useMe=TRUE)
sub2 <- list(data=df2, useMe=TRUE)
sub3 <- list(data=df3, useMe=FALSE)

#Store the sublists in a main list
main <- list(sub1, sub2, sub3)

I want a vectorized function that will return the row-wise average of the data frames, but only if $useMe==TRUE, like so:

> desiredFun(main)
   id  data
1   a     1
2   b     2
3   c     3

Answer 1

Here’s a fairly general way to approach this problem:

# Extract the "data" portion of each "main" list element
# (using lapply to return a list)
AllData <- lapply(main, "[[", "data")
# Extract the "useMe" portion of each "main" list element
# using sapply to return a vector)
UseMe <- sapply(main, "[[", "useMe")
# Select the "data" list elements where the "useMe" vector elements are TRUE
# and rbind all the data.frames together
Data <- do.call(rbind, AllData[UseMe])
library(plyr)
# Aggregate the resulting data.frame
Avg <- ddply(Data, "id", summarize, data=mean(data))