Open TylerGrantSmith opened 5 months ago
Can reproduce.
If it helps with debugging: it runs when projection_pushdown
is disabled.
>>> print(pl.collect_all([ldf.select(COLUMNS[0])], projection_pushdown=False))
[shape: (1, 1)
┌─────┐
│ 0 │
│ --- │
│ str │
╞═════╡
│ a │
└─────┘]
Can reproduce and projection_pushdown=False
only seems to make a higher value for NVARS
possible for me before segfault
You segfault because we StackOverFlow. This happens at a certain NVARS
.
I’m facing the same issue, and it seems to happen also with collect
. I’ll try to work on a reproduction.
Thanks all, it looks like the issue was resolved along with many others as of 0.20.17
Pre-mature....the example passed, but still overflows (at ~N=500) now
Checks
Reproducible example
Log output
Issue description
This is a reprex distilled from a more complicated internal process. The value of
NVARS
at which the errors start varies by machine. On my current machine the value decreased going from 0.20.4 to 0.20.6Expected behavior
To run without segfaulting.
Installed versions