How to sum each row in a .xts object, where values are NOT missing

Question

I currently have a line of code set up to sum each row/date of a xts object, where value for each data point = 1:

   universe.rt=sapply(X=2:nrow(rt),FUN=function(x){sum(rt[x,which(live[x,]==1)])/count[x]})

I want to change the code such that instead of summing up all points in a row where value = 1, I want to sum all points in a row where value is NOT NA:

   universe.rt=sapply(X=2:nrow(rt),FUN=function(x){sum(rt[x,which(live[x,]==!is.na)])/count[x]})

I can't get the syntax right.

score 4 · Answer 1 · answered May 27 '15 at 11:22

4

for xts objects, you should use rowSums rather than sapply, because the apply functions use loops internally, and rowSums calls out to faster C code.

rowSums(x,na.rm=TRUE)

answered May 27 '15 at 11:22

Brian G. Peterson

3,531
2
20
21

Thanks for the additional input Brian - it is much faster now! – Malik Jun 04 '15 at 14:26

score 0 · Accepted Answer · answered May 26 '15 at 20:20

0

is.na is a function, you should use

which(!is.na(live[x,]))

answered May 26 '15 at 20:20

cyberj0g

3,707
1
19
34

How to sum each row in a .xts object, where values are NOT missing

2 Answers2