r - variable values in data.table::melt with patterns

Question

While answering this question I tried to use the patterns functionality together with data.table::melt:

df1 <- structure(list(Material_code = 111:112, 
                      actual_202009 = c(30L, 19L), 
                      actual_202010 = c(44L, 70L), 
                      actual_202011 = c(24L, 93L), 
                      pred_202009 = c(25L, 23L), 
                      pred_202010 = c(52L, 68L), 
                      pred_202011 = c(27L, 100L)), 
                 class = c("data.table", "data.frame"), 
                 row.names = c(NA, -2L))

I wanted to convert that wide table to a long one, separating the actual and pred values. I thought the date part was going to be conserved in the variable, but it was changed to factor numbers (1, 2, etc):

melt(df1, 1, measure = patterns(actual = "actual_", pred = "pred_"))[1, ]
>    Material_code variable actual pred
> 1:           111        1     30   25

I wanted "202009" instead of "1" in the variable field.

I am fully aware I can achieve it the long way, using tstrsplit:

melt(df1, 1)[, 
           c("type", "variable") := tstrsplit(variable, "_", fixed = TRUE)][,
                 dcast(.SD, Material_code + date ~ type)]

but I was expecting that with patterns I could produce a less verbose way.

What I've tried:

• I changed variable.factor = FALSE (default is true), but it produces the same values in the variable column (1, 2, ...), only this time they are strings.

Expected output:

I expect the following code to produce the shown output (of course replacing argument1 and argument2 with whatever is appropriate):

melt(df1, 1, measure = patterns(actual = "actual_", pred = "pred_"), argument1, argument2)[1, ]
>    Material_code variable actual pred
> 1:           111   202009     30   25

Is it possible, or definitely one has to go the long way?

question from:https://stackoverflow.com/questions/65937422/variable-values-in-data-tablemelt-with-patterns

Answer 1

Waitting for answers