Query on parquet files hangs in 1.1+ but not in 1.0 #42

What happens?

I'm still in the process of paring down my query into a minimal repro (involving hundreds lines of duckdb SQL -- mostly the same column names repeated in every subquery -- and 1M row parquet files) but wanted to flag this in case there's a known issue I'm missing: when running a complicated query that entails joining two parquet tables, the query hangs indefinitely in versions 1.1 onward; however, in 1.0 it finishes in seconds. Whenever I enable trace logging, the last output is

[LOG] 2025-09-10 04:38:12.669, FileSystem, TRACE, {"fs":"LocalFileSystem","path":"hanging_2/orderlog_X1t_20250901.current.parquet","op":"READ","bytes":"22","pos":"5782417"}, CONNECTION, 2, 3, NULL

I've attempted to bisect the parquet file (by trimming off large sections off it) but it seems like the error is non-deterministic. Is it possible to see if there's a problematic block / value associated with the pos field of 5782417?

What I do know is that this doesn't seem specific to the query plan; if I trim off enough of the file, it finishes instantly.

To Reproduce

Still working on getting a repro example

OS:

RHEL 8.6

DuckDB Package Version:

1.1+ (tested in 1.3.2)

Python Version:

3.12

Full Name:

Zach Silversmith

Affiliation:

None

What is the latest build you tested with? If possible, we recommend testing with the latest nightly build.

I have tested with a stable release

Did you include all relevant data sets for reproducing the issue?

No - Other reason (please specify in the issue body)

Did you include all code required to reproduce the issue?

• Yes, I have

Did you include all relevant configuration to reproduce the issue?

• Yes, I have