How do you pivot data from a list of data frames in R?

Let's say I have a list of data frames ldf:

df1 <- data.frame(date = c(1,2), value = c(4,5))
df2 <- data.frame(date = c(1,2), value = c(4,5))
ldf <- list(df1, df2)

What is the best way to get the sum (or any other function) of values by date, i.e. some data frame like:

data.frame(date = c(1,2), value = c(8,10))

Answer #1:

You could use:

library(data.table)
dt1 <- rbindlist(ldf)
setkey(dt1,'date')
dt1[,list(value=sum(value)), by='date']
 date value
1:    1     8
2:    2    10
Answered By: akrun

Answer #2:

If these rows were all in the same data frame, you would use aggregate to do the sum. You can combine them with rbind so they are in the same data frame:

aggregate(value ~ date, data=do.call(rbind, ldf), FUN=sum)
  date value
1    1     8
2    2    10

If the date columns in all the data frames are identical, you can easily use Reduce to do the sum:

Reduce(function(x, y) data.frame(date=x$date, value=x$value+y$value), ldf)
  date value
1    1     8
2    2    10

This should be a lot faster than rbind-ing the data together and aggregating.

Answered By: Matthew Lundberg

Answer #3:

Another option is to use unnest from "tidyr" in conjunction with the typical grouping and aggregation functions via "dplyr":

library(dplyr)
library(tidyr)

unnest(ldf) %>%
  group_by(date) %>%
  summarise(value = sum(value))
# Source: local data frame [2 x 2]
# 
#   date value
# 1    1     8
# 2    2    10
The answers/resolutions are collected from stackoverflow, are licensed under cc by-sa 2.5 , cc by-sa 3.0 and cc by-sa 4.0 .



# More Articles