Transform a matrix 'multiplying' each row name by column name in R

Question

I need to do a pretty common transformation, but I forgot how it is called. Basically I have a matrix:

And I want to get a different matrix with each row to be: gene_id, Vx, num. For instance:

WBGene00000001, V1, 0
WBGene00000002, V2, 0
...

I have a dataframe in R, so how could it be approached there?

Sorry, but I need not only diagonal elements. If I just supply `mat`, would it work? — Nikita Vlasenko, Jan 31 '18 at 01:10
many rows like that: (WBGene00000006, Vx, num), a combination of WBGene00000006 with all of the Vs — Nikita Vlasenko, Jan 31 '18 at 01:17
`expand.grid(row.names(mat),colnames(mat),c(mat))`. If it is a dataframe then you will use `unlist` instead of `c` — Onyambu, Jan 31 '18 at 01:19

score 2 · Accepted Answer · answered Jan 31 '18 at 01:23

This sounds like a pivot/unpivot operation. Easiest implementation is via tidyverse::gather() function:

x = readr::read_csv("gene_id,V1,V2\nWBGene00000001,0,0\nWBGene00000002,0,0")
x

         gene_id    V1    V2
           <chr> <int> <int>
1 WBGene00000001     0     0
2 WBGene00000002     0     0

tidyr::gather(x, varname, value, V1:V2)

         gene_id varname value
           <chr>   <chr> <int>
1 WBGene00000001      V1     0
2 WBGene00000002      V1     0
3 WBGene00000001      V2     0
4 WBGene00000002      V2     0

Transform a matrix 'multiplying' each row name by column name in R

1 Answers1