R data.table binary value for last row in group by condition

For each id, check if row number is the last row number in the group, and if 'conversion' is 1. Convert logical result to integer.

DT[ , lastconv := as.integer(.I == .I[.N] & conversion == 1), by = id]

Modifying the OP's code to join on the last row of each group:

DT[, v := 0]
DT[.(DT[conversion == 1, unique(id)]), on=.(id), mult="last", v := 1]

     id conversion v
1: 1232          0 0
2: 1232          0 0
3: 1232          0 0
4: 4211          1 0
5: 4211          1 0
6: 4211          1 1

This is only different in that it selects which ids to edit based on the desired condition.

Filter for the last row per group and set lastconv equal to conversion.

DT[DT[, .I[.N], by=id]$V1, lastconv := conversion]

Then replace NAs with 0

DT[is.na(lastconv), lastconv := 0L]

Result

DT
#     id conversion lastconv
#1: 1232          0        0
#2: 1232          0        0
#3: 1232          0        0
#4: 4211          1        0
#5: 4211          1        0
#6: 4211          1        1

---

If data.table v1.12.3 is installed we could also use the new function setnafill to replace NAs in the second step

DT[DT[, .I[.N], by=id]$V1, lastconv := conversion]
setnafill(DT, cols = "lastconv", fill = 0L)