Question

I would like to state the row/col output names in a **ply function, ldply, from the plyr package.

for example,

I have a list, foo, that I want to convert to a data.frame and truncate significant digits with signif()

 foo <- list(var.a = runif(3), var.b = runif(3), var.c=runif(3))

What I have now is

q <- ldply(foo, signif, 2)
colnames(dq)[1] <- c('id', 'q1', 'q2','q3')
rownames(dq) <- dq$id

Is there an easier way?

Two previous questions have asked how to use plyr to rename rows and cols using plyr, but I think my question is different. Can the names be stated at the same time as another function (or if I am doing this correctly)? Is this a worthwhile feature request?

Answer 1

You have to give names somewhere, either on the function being called inside as eg in

R> ldply(foo, function(l) c(a=signif(l[1], 2), b=signif(l[2], 2), 
+                           c=signif(l[3], 2)))
    .id    a    b    c
1 var.a 0.50 0.72 0.27
2 var.b 0.82 0.38 0.24
3 var.c 0.13 0.27 0.81
R> 

or has you did after the call.

Another option I frequently use is to explicitly create one-row data.frame in the anonymous worker function. *dply() et al then simply collated these into a single data.frame. That works well enough for my tastes.