Remove rows in data.table according to another data.table

Question

I have a data.table named dtA:

My actual dtA has 62871932 rows and 3 columns:

  date    company    value
19810         


        

Answer 1

I think I know how to solve this:

in dtB I add a pointer using data.table syntax:

dtB[, pointer := 1]

dtB will looks like this

  date    company    value    pointer
198101          A        2          1
198102          B        5          1

Then I use LEFT OUTER JOIN method from here: https://rstudio-pubs-static.s3.amazonaws.com/52230_5ae0d25125b544caab32f75f0360e775.html

setkey(dtA, date, company, value)
setkey(dtB, date, company, value)
dtA=merge(dtA, dtB, all.x)

This means on pointer column, if dtB's row exist in dtA, it will give 1. If dtB's row do not exist in dtA's, then it will be given NA

Result will be:

  date    company    value    pointer
198101          A        1         NA
198101          A        2          1
198101          B        5         NA
198102          A        2         NA
198102          B        5          1
198102          B        6         NA

I then select those rows with NA and remove pointer column:

dtA=dtA[!(pointer %in% "1")][,-c("pointer")]

I get my result:

  date    company    value
198101          A        1
198101          B        5
198102          A        2
198102          B        6