Pass parameter to lapply as variable name in data.table

Question

I'm trying to pass an argument to a function using lapply in datatable, only the argument is a variable name (It is called in a function). The wtd.var function is from Hmisc. Example using iris:

library(Hmisc)
library(data.table)
iris <- as.data.table(iris)
weights.v ="Sepal.Length" 
iris[,lapply(.SD,wtd.var,weights=weights.v),by = "Species"]

throws an error, naturally since because the variable is a string. I tried various combinations or paste0 and parse, but get() seems to do it-

iris[,lapply(.SD,wtd.var,weights=get(weights.v)),by = "Species"]

That works. But when I try to use .SDcols, it gets ignored

iris[,lapply(.SD,wtd.var,weights=weights.v),by = "Species",.SDcols=c("Petal.Length")]

Is there a better way to achieve what I'm doing?

@DavidArenburg, and on my session it did seem to work - why shouldn't it? — talat, May 06 '15 at 11:49
@docendodiscimus because of [this](http://stackoverflow.com/questions/23361080/error-in-setdt-from-data-table-package) it is very strange it worked for you. It doesn't work for me and this is documented behaviour (as per Aruns answer). — David Arenburg, May 06 '15 at 11:51
@DavidArenburg, well, I don't any single error message and iris prints just fine (as data.table) afterwards.. — talat, May 06 '15 at 11:52
@docendodiscimus very odd. It should only work if you already created a copy of it once. If you'll try it on a fresh session is still works? — David Arenburg, May 06 '15 at 11:54
This works: `iris[,lapply(.SD[,"Petal.Length",with=FALSE],wtd.var,weights=get(weights.v)),by = "Species"]` — Frank, May 06 '15 at 14:46

score 0 · Answer 1 · answered May 06 '15 at 12:45

0

A simple (but probably sub-optimal) workaround would be to sapply by subsets...

sapply(levels(iris$Species), function(x) {with(subset(iris,
Species==x),wtd.var(Petal.Length,Sepal.Length))})

Hope this helps.

answered May 06 '15 at 12:45

GPP

46
6

Pass parameter to lapply as variable name in data.table

1 Answers1