Making some columns if a dataframe numeric

Question

I have two dataframes

> head(a[1:4,1:4])
   Tumor_Sample_Barcode Chromosome Position End_Position
1:                   A1       chr4 90169866     90169866
2:                   A1      chr11 60235747     60235747
3:                   A1       chr1   983023       983023
4:                   A1       chr1 11346060     11346060
> 

> head(c)
   Chromosome Position      VAF
1:       chrM     6691 0.610284
2:       chrM    14503 0.693325
3:       chr1 31412236 0.645161
4:       chr1 55693305 0.602941
5:       chr1 69963412 0.709302
6:       chr1 72608266 0.720000
>

> str(c)
Classes ‘data.table’ and 'data.frame':  44175 obs. of  3 variables:
 $ Chromosome: chr  "chrM" "chrM" "chr1" "chr1" ...
 $ Position  : num  6691 14503 31412236 55693305 69963412 ...
 $ VAF       : num  0.61 0.693 0.645 0.603 0.709 ...
 - attr(*, ".internal.selfref")=<externalptr> 

I want to merge them like this but complaining about not being integer

> a=a[c, on = c("Chromosome","Position"), VAF := i.VAF]
Error in bmerge(i, x, leftcols, rightcols, roll, rollends, nomatch, mult,  : 
  Incompatible join types: x.Position (character) and i.Position (integer)

I have tried something like these but not working

> c=c[!is.na(as.numeric(as.character(c$Position))), ]
> 
> c=c[!is.na(as.numeric(as.character(c$VAF))), ]
> 
> c$VAF <- as.numeric(as.character(c$VAF)) 

c$Position <- as.numeric(as.character(c$Position)) 

which not working

Answer 1

Your issue seems to be related to the column Position being a character column in a, you should try:

library(data.table)
a$Position <- as.numeric(a$Position)
a[c, on = c("Chromosome","Position"), VAF := i.VAF]