[datatable-help] Linux versus Windows behavior

[Matthew Dowle]

Hi,

Can't say I've seen that error exactly, but similar. For the difference
between machines, does test.data.table() return precisely 717 tests passed
on the Windows machines?

If so, there are some bug fixes in that area in 1.8.3 :

o   In X[Y,cols,with=FALSE], NA matches are now handled correctly. And if
cols includes join columns, NA matches (if any) are now populated from i.
For consistency with X[Y] and X[Y,list(...)]. Tests added.

o   "Internal error" message when combining join containing missing groups
and group by is fixed, #2162. For example :
  X[Y,.N,by=NonJoinColumn]
where Y contains some rows that don't match to X. This bug could also
result in a seg fault. Thanks to Andrey Riabushenko and Michael
Schermerhorn for reporting. Tests added.

Perhaps upgrade if possible, can always revert back. A memory corruption
can often not show up on one machine vs another.

Matthew


> All,
>
> This refers to data.table 1.8.2 on R 2.15.1
>
> I'm doing a large join using ':=' on some tables and I have no problem
> doing the join on a Linux box but on Windows (my 32bit laptop as well as a
> colleagues 64 bit windows)
> we keep getting the following error.
>
> Error in `[.data.table`(slot.data, tmp.data[, key(slot.data), with =
> FALSE],  :
>   Internal error: NA exist in 'rows' passed to C assign
>
> Has anyone else witnessed similar behavior?
>
> Best regards,
>
> Damian
>
>
> _______________________________________________
> datatable-help mailing list
> datatable-help at lists.r-forge.r-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/datatable-help