Invalid row order after join

[mllg]

MWE:

library(data.table)

A = data.table(i = 1:6, j = rep(1:2, 3), x = letters[1:6], key = "i")
B = data.table(j = 1:2, y = letters[1:2], key = "j")

res = A[B, on = "j"]

res has key i but is unordered:

   i j x y
1: 1 1 a a
2: 3 1 c a
3: 5 1 e a
4: 2 2 b b
5: 4 2 d b
6: 6 2 f b

Setting the key manually (setkeyv(res, "i")) yields a warning.

Session info:

R version 3.3.1 (2016-06-21)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Arch Linux

locale:
 [1] LC_CTYPE=de_DE.UTF-8       LC_NUMERIC=C               LC_TIME=de_DE.UTF-8        LC_COLLATE=C               LC_MONETARY=de_DE.UTF-8    LC_MESSAGES=en_US.UTF-8    LC_PAPER=de_DE.UTF-8       LC_NAME=C
 [9] LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=de_DE.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

other attached packages:
[1] data.table_1.9.6 nvimcom_0.9-19   setwidth_1.0-4

loaded via a namespace (and not attached):
[1] tools_3.3.1    parallel_3.3.1 chron_2.3-47

[mllg]

Side note: You can still index the resulting table, but the lookup on i just does not work. E.g.,

res[.(2), nomatch = 0]

returns a data.table with 0 rows.

[jangorecki]

Thanks for reporting. Reproducible on 1.9.7. The actual warning:

In setkeyv(x, cols, verbose = verbose, physical = physical) :
  Already keyed by this key but had invalid row order, key rebuilt. If you didn't go under the hood please let datatable-help know so the root cause can be fixed.

Looks like key is not removed during join, it should be removed. Optionally results could have key set on j column, as the key of B.