Question

I am looking to take data that is currently wide and melt it into a long format. The trick is that I want to create a sequence indicator.

Here are my data:

df.wide <- data.frame(id = 1:5, 
                      code1 = sample(month.abb, 5),
                      code2 = sample(month.abb, 5))

What I am looking for:

   id rank value
1   1    1       Dec
2   1    2       Jan
3   2    1       May
4   2    2       Jun
5   3    1       Aug
6   3    2       Aug
7   4    1       Sep
8   4    2       Mar
9   5    1       Dec
10  5    2       Nov

I suspect that I could use this:

melt(df.wide, id=c("id"))

and iterate over the data aftewards to clean up the results, but I know reshape is a great package and wanted to ask before I reinvent the wheel.

Answer 1

Here is a one-liner using the reshape function in base R (not to be confused with the reshape package)

reshape(df.wide, varying = 2:3, timevar = 'rank', sep = "", direction = 'long')