-
Notifications
You must be signed in to change notification settings - Fork 27
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Undefined behavior with list columns #141
Comments
Also an issue with R 4.3.3. |
Running with valgrind:
|
Another one, very similar:
@hannes: Have you recently seen undefined behavior with persisting nested data? I'll check the latest and greatest too. |
Still an issue with 953e7d3 (aka duckdb/duckdb@e5f4523), installed from r-universe:
|
Can't repro with the CLI: CREATE TABLE test_list_2 (a integer, b STRUCT(c VARCHAR[], d VARCHAR[], e INTEGER[]));
INSERT INTO test_list_2 VALUES (1, row(['a', 'b', 'c', 'd', 'e', 'f'], ['A', 'B'], [1, 5, 9]));
SELECT * FROM test_list_2; |
But with
I'll open an issue upstream. |
@TSchiefer: can you confirm the current development version (installable as a binary from r-universe) fixes the issue? |
I installed package version 0.10.1.9001 from r-universe today (I see that version bump is from 3 weeks ago), but it was installed from source, not entirely sure that it's the correct version & error is unfortunately still occurring. |
it might be intentional, AFAIK we don't zero blocks for performance reasons and if there's garbage at the end it just gets written to disk. But need to confirm. |
@TSchiefer: can you please share a reproducible example for the issues you are seeing now? |
Thanks, I received an example. The original test still fails with v0.10.3, testing against bleeding-edge duckdb now. |
I can confirm that even the original issue still persists after vendoring the most recent duckdb (duckdb/duckdb@c3903e4). I'll bring this upstream again. |
To reproduce:
docker run --rm -ti --platform linux/amd64 -v $(pwd):/test ghcr.io/krlmlr/r-4-duckdb:main R -q -f test/test.R
Symptoms:
The number varies, it is high on gitpod and on another Linux host (bare metal), hundreds of MB on macOS + colima, the problem can't be reproduced with Docker Desktop on macOS.
Need to understand if this is specific to R 4.0.5: krlmlr/r-4-duckdb#1.
The text was updated successfully, but these errors were encountered: