Reputation: 400
I can't make chaining work in this example. Can someone explain what I was missing?
library(data.table)
dt <- data.table(a=c(rep("komm", 5), rep("by", 5)), paste0("nr.",1:10))
dt[a=="komm", v3:=sub("nr.", "", V2)]
dt[, v4:=sub("\\D*(\\d)", "\\1", V2)]
# doesn't work
dt[a=="by"][
, v5:=sub("nr.", "no.", V2)][
, v6:=sub("\\D*(\\d)", "\\1", V2)]
I was expecting to get this output
a V2 v3 v4 v5 v6
1: komm nr.1 1 1 <NA> <NA>
2: komm nr.2 2 2 <NA> <NA>
3: komm nr.3 3 3 <NA> <NA>
4: komm nr.4 4 4 <NA> <NA>
5: komm nr.5 5 5 <NA> <NA>
6: by nr.6 <NA> 6 no.6 6
7: by nr.7 <NA> 7 no.7 7
8: by nr.8 <NA> 8 no.8 8
9: by nr.9 <NA> 9 no.9 9
10: by nr.10 <NA> 10 no.10 10
Upvotes: 2
Views: 170
Reputation: 160687
Once you filter and close the bracket, the in-place assignment is broken. That is, DT[cond,newvar:=1]
assigns based on a condition, and is equivalent to DT[,newvar:=fifelse(cond,1,newvar)]
or similar. However, DT[cond,]
is returning a new frame, and any work on it now is completely separate from the original DT
.
Either do the conditional assignment twice
dt <- data.table(a=c(rep("komm", 5), rep("by", 5)), paste0("nr.",1:10))
dt[a=="by", v3 := sub("nr.", "", V2)][a=="by", v4:=sub("\\D*(\\d)", "\\1", V2)]
# a V2 v3 v4
# <char> <char> <char> <char>
# 1: komm nr.1 <NA> <NA>
# 2: komm nr.2 <NA> <NA>
# 3: komm nr.3 <NA> <NA>
# 4: komm nr.4 <NA> <NA>
# 5: komm nr.5 <NA> <NA>
# 6: by nr.6 6 6
# 7: by nr.7 7 7
# 8: by nr.8 8 8
# 9: by nr.9 9 9
# 10: by nr.10 10 10
... or once with multi-assignment:
dt <- data.table(a=c(rep("komm", 5), rep("by", 5)), paste0("nr.",1:10))
dt[a=="by", c("v3", "v4") := .(sub("nr.", "", V2), sub("\\D*(\\d)", "\\1", V2))]
# a V2 v3 v4
# <char> <char> <char> <char>
# 1: komm nr.1 <NA> <NA>
# 2: komm nr.2 <NA> <NA>
# 3: komm nr.3 <NA> <NA>
# 4: komm nr.4 <NA> <NA>
# 5: komm nr.5 <NA> <NA>
# 6: by nr.6 6 6
# 7: by nr.7 7 7
# 8: by nr.8 8 8
# 9: by nr.9 9 9
# 10: by nr.10 10 10
Upvotes: 2