I'm working with a system that outputs data in "R dump" format. For example it might output a three-dimensional array looking like this:
obs <- structure(c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24),
.Dim=c(2,4,3))
I'm new to R but I would like to use R to inspect marginal summaries of this data. For example, I'd like to see a 2x4 table of mean values averaged across that third dimension.
(If possible, I'd also like to see marginal summaries collapsed down to one dimension, eg a row of 4 mean values, each mean being taken over a 2x3 slice of my data.)
I tried summary(obs)
which collapses all the dimensions and gives overall stats, and sapply(obs, summary)
which doesn't collapse any of the dimensions, just giving a "summary" of each individual datum.
I expect there's a function for what I'm after but I can't find it!
apply
works for this:
apply(obs,1:2,mean)
[,1] [,2] [,3] [,4]
[1,] 9 11 13 15
[2,] 10 12 14 16
or
aperm(apply(obs,1:2,summary),c(1,3,2))
(or apply(obs,2:1,summary)
as pointed out in comments)
with results:
[,1] [,2] [,3] [,4]
Min. 1 3 5 7
1st Qu. 5 7 9 11
Median 9 11 13 15
Mean 9 11 13 15
3rd Qu. 13 15 17 19
Max. 17 19 21 23
, , 2
[,1] [,2] [,3] [,4]
Min. 2 4 6 8
1st Qu. 6 8 10 12
Median 10 12 14 16
Mean 10 12 14 16
3rd Qu. 14 16 18 20
Max. 18 20 22 24
As requested you can get other marginal summaries
apply(obs,2,mean)
## [1] 9.5 11.5 13.5 15.5
(double-check: mean(obs[,1,])
is indeed 9.5 ...)
While digging in the R toolbox, you may also wish to check the plyr
tool: a*ply
. The function takes an array
as input, and it is easy to control in which form the result is returned: array, data frame or list.
To make it a little bit easier to keep track of the dimensions when we play around with your example array, I added some arbitrary dimension names. The first dimension (rows) = species; second (columns) = time; third (separate 'tables') = site
obs <- array(c(1:24),
dim = c(2, 4, 3),
dimnames = list(species = c("cat", "dog"),
time = 1:4,
site = letters[1:3]))
library(plyr)
# result as (2-d) array: aaply
# i.e. same output as @Ben Bolker's `apply` example
# keep the first two dimensions (species, time), collapse the third (site)
aaply(obs, 1:2, mean)
# time
# species 1 2 3 4
# cat 9 11 13 15
# dog 10 12 14 16
# result as data frame: adply
adply(obs, 1:2, mean)
# species time V1
# 1 cat 1 9
# 2 dog 1 10
# 3 cat 2 11
# 4 dog 2 12
# 5 cat 3 13
# 6 dog 3 14
# 7 cat 4 15
# 8 dog 4 16
# several functions
adply(obs, 1:2, each(min, mean, max))
# species time min mean max
# 1 cat 1 1 9 17
# 2 dog 1 2 10 18
# 3 cat 2 3 11 19
# 4 dog 2 4 12 20
# 5 cat 3 5 13 21
# 6 dog 3 6 14 22
# 7 cat 4 7 15 23
# 8 dog 4 8 16 24
# apparently the `each` thing can be used on just one function as well,
# then the function name appears as column name instead of 'V1' as above.
adply(obs, 1:2, each(mean))
# species time mean
# 1 cat 1 9
# 2 dog 1 10
# 3 cat 2 11
# 4 dog 2 12
# 5 cat 3 13
# 6 dog 3 14
# 7 cat 4 15
# 8 dog 4 16
# one-dimensional summary
adply(obs, 2, each(mean))
# time mean
# 1 1 9.5
# 2 2 11.5
# 3 3 13.5
# 4 4 15.5
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.