Subsetting data.table can be done using base R subset with S3 Class for
data.table object. Alternatively is to use the recode approach as mentioned in
StackOverflow.
lookup <- list(v1 = 1:3, v2 = letters[5:7])
DT[lookup, on = names(lookup), nomatch = NULL]
This will return all columns in DT that match lookup list. Specifying with
nomatch = NULL or nomatch = 0 will exclude rows in DT that don’t match
lookup.
Another way to do dynamic subset is as below, but it’s a bit slower compared to the recode approach:
DT[DT[, all(mapply(`%in%`, .SD, lookup)), by = 1:nrow(DT), .SDcols=names(lookup)]$V1]
A nice reading on table join for data.table which is very relevant to this case is the blog post by Scott Lyden.