Selecting columns based on certain conditions in a data.table in R

Question

I have a data.table in R with 200 columns with integer values.

One of the columns is named group and it has 100 different possible values.

So, when I subset using: subDT<-DT[group==N], for instance, and if I do, sum(subDT$columnX), maybe the value of the sum will be 0.

So what I want to do is to display subDT, but only the columns where sum(subDT$columnN)!=0, something like subDT[group==0,.(columns where sum(column)>0)], keeping the names of the columns intact of course.

EDIT

An example using the mtcars data would be:

DT<-as.data.table(mtcars)

Let's say that we want to subset mtcars and get the rows where carb is 1, but display the columns only if the sum of the subset is less than 10:

DT[carb == 1, (sapply(DT[carb == 1],sum) < 10), with = FALSE]

In this case, the columns that will be displayed are only vs,am and carb because the sum of those columns is less than 10

Please provide a [reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) of your data. Is it really a `data.table` or a `data.frame`? — Molx, Jul 09 '15 at 18:27
You should really provide a reproducible example, because as it stands now, this questions isn't useful to no one and should be closed IMO — David Arenburg, Jul 09 '15 at 18:47
I really made my best effort to be as clear as possible including data.table in the title and tag and being as explicit as possible... Thanks anyway the solution is down there — Felipe, Jul 09 '15 at 18:48
Did you follow [this](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) link or not? — David Arenburg, Jul 09 '15 at 18:51
I did... I added the reproducible example so maybe it can be more useful to people. — Felipe, Jul 09 '15 at 19:04
@akrun probably `lapply(.SD, sum) < 10` will be more efficient. — David Arenburg, Jul 09 '15 at 19:11

score 3 · Accepted Answer · answered Jul 09 '15 at 18:32

3

Assuming that the sum is taken after the filter:

DT[group == N, !(sapply(DT[group == N],sum) == 0), with = FALSE]

Can be made faster with setkey:

setkey(DT,group)
DT[N, !(sapply(DT[N],sum) == 0), with = FALSE]

answered Jul 09 '15 at 18:32

Chris

6,302
1
27
54

Selecting columns based on certain conditions in a data.table in R

1 Answers1