I have a wide R dataframe that I want to partially melt, where the columns all share the same sets of time, and the syntax of the column-names is of a the form (varname_tt_(integer)). Is there a nice way to do the transform below using tidyverse and pivot_longer()? If not, is there some alternative way that's fairly clean?
Here is reproducible example, where I want to transform df into df_partial_melt
df = data.frame(
name=c("A","B","C"),
XX_tt_1 = c(10,20,30),
XX_tt_2 = c(100,200,300),
YY_tt_1 = c(40, 50, 60),
YY_tt_2 = c(400, 500, 600)
)
print(df)
name XX_tt_1 XX_tt_2 YY_tt_1 YY_tt_2
1 A 10 100 40 400
2 B 20 200 50 500
3 C 30 300 60 600
df_partial_melt = data.frame(
name=rep(c("A","B","C"), each = 2),
time = rep(c(1,2), 3),
XX = c(10, 100, 20, 200, 30, 300),
YY = c(40,400,50,500,60,600)
)
> print(df_partial_melt)
name time XX YY
1 A 1 10 40
2 A 2 100 400
3 B 1 20 50
4 B 2 200 500
5 C 1 30 60
6 C 2 300 600
