r - data.table: merge() one-key table with two-keys table -

i refer this questions 3 years old. have similar problem.

i have these 3 tables:

require(data.table)  profile <- data.table(id = seq(11,20)) setkey(profile, "id")  post <- data.table(id = seq(1,10)) setkey(post, "id")  comment <- data.table(post_id = seq(1,10),                       profile_id = seq(11,20)) setkeyv(comment, c("post_id","profile_id")) 

now want merge profile , post comment 2 different tables. how specify key match in comment profile_id , in post post_id? should respecify how tables built?

unfortunately, merge by.x= by.y= in merge data.frame not implemented data.tables yet. adressed in next release 1.9.6. see here: https://github.com/rdatatable/data.table/issues/637.

what can convert data.tables data.frames, merge there using by.x , by.y (see ?merge.data.frame that) , convert data.tables.

alternatively, name keys/columns in way match on name basis. then, data.table's merge should work.

setnames(profile, "id", "profile_id") setnames(post,    "id", "post_id")  merged_dt1 <- merge(profile, comment) merged_dt2 <- merge(post, comment)