CODEWITHSUNDEEP
rlist
raK1
raK1

Reputation: 529

fastest way to add elements in list in R

zii=list()
zii[[1]]=c(1,2,3)
zii[[2]]=c(1,2,3)
zii[[3]]=c(1,2,3)

What is the best way to perform element-wise addition in the list , IE: sum=c(1+1+1,2+2+2,3+3+3)=c(3,6,9) I tried Reduce("+",zii) and it is slow. Any other suggestions ?

Upvotes: 2

Views: 120

Answers (1)

IRTFM
IRTFM

Reputation: 263382

I'm not sure whether this will be any faster. The data.frame does a lot of validity checking:

> rowSums(data.frame(zii))
[1] 3 6 9

Could also try these if you get around to using microbenchmark. I'm guessing one of these will win and my money would be on the second one.:

> rowSums(do.call(cbind, zii))
[1] 3 6 9
> colSums(do.call(rbind, zii))
[1] 3 6 9

Looks like I lost my bet:

require(microbenchmark)
microbenchmark(   Reduce("+",zii) , 
                  rowSums(data.frame(zii)),
                  rowSums(do.call(cbind, zii)),
                  colSums(do.call(rbind, zii)) )
#------------------------------------------------------
Unit: microseconds
                         expr     min       lq      mean   median       uq
             Reduce("+", zii)  26.975  28.1870  31.02119  30.0560  30.9695
     rowSums(data.frame(zii)) 730.933 744.9015 776.36775 753.5785 787.2765
 rowSums(do.call(cbind, zii))  65.770  67.3800  71.94039  68.7050  70.1335
 colSums(do.call(rbind, zii))  61.202  62.8830  66.21362  64.1060  65.9130
      max neval cld
   57.958   100 a  
 1129.953   100   c
  176.627   100  b 
  127.259   100  b

Upvotes: 4

Related Questions