feature request : transpose(keep.rownames = T)?

[ghost]

transposing up to now required this:

dt = data.table(x=1:5, y=6:10)
t(dt) %>% data.table(keep.rownames = T)

, so the transform function has potential to cut this short.
It has however no keep.rownames = T argument (or something similar), so the variable names disappear:

transpose(dt)
#    V1 V2 V3 V4 V5
#1:  1  2  3  4  5
#2:  6  7  8  9 10

what I hoped to see was something like this:

#    rn  V1 V2 V3 V4 V5
#1:  x  1  2  3  4  5
#2:  y  6  7  8  9 10

[skanskan]

This should be easy to implement and an important addition.

[MichaelChirico]

could you elaborate with sample input & output? lists don't really have a "row names" attribute...

…

On Mar 8, 2017 6:45 AM, "skanskan" ***@***.***> wrote: This should be easy to implement and an important addition. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1886 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AHQQdT_v8cy-pMpmR4pTa2uAN090SEZDks5rjpTagaJpZM4KcVr3> .

[skanskan]

Imagine you have a data.table with several columns, such as this toy example.
dt = data.table(A=1:5, B=c(NA,3,3,NA,NA), C=c(NA,NA,9,8,7))

and you want to calculate several things to summarize the different variables.

nas <- t(dt[,lapply(.SD,function(x) sum(is.na(x)))])
sums  <- t(dt[,lapply(.SD,function(x) sum(x,na.rm = T))])
data.table(nas=nas,sums=sums)

Something similar to this:
t(apply(dt,2,function(x) rbind(sum(x,na.rm = T),sum(is.na(x)))))

But I having a result with colnames and rownames:

   nas   sums
A   15    0
B    6    3
C   24    2

[ghost]

if the example input & output in the original feature request is not clear enough let me know and I will edit.

I agree that a list does not have row.names. The function is also intended for data.frames and data.tables however. When you transpose these objects, their variable names would become row.names, in which the transpose function does not foresee. Hence the need for the verbose as.data.table(t(dt), keep.rownames = T).

[MichaelChirico]

This could have been useful in this SO post -- instead of needing the helper function coefsplit to break apart the named result of coef(...), we could do transpose(as.list(coef(...)), keep.rownames = TRUE).

This comes at a sort of general thing that I run into a lot of needing to split the named result of a routine into separate columns because the names are important (most salient and consistent example being when I run table by group)