Open JAicewizard opened 1 month ago
Thanks for the bug report! Ill try to have a look sometime next week.
Updated the script to use only 2 dimensions, error still reproduces. Reducing to just one and the error disapears, but this should make debugging a bit easier
I was running this extension on my dataset, and I got this error:
Could not find node in column segment tree! Attempting to find row number "11891708928" in 2595 nodes
followed by a bunch of nodes and some info about them.This is reproducible on vss as of 11/5/2024 18:24 (dutch time) on both 0.10.1 and 0.10.2. The version of duckdb is self-build (more specifically, build as part of building an extension, in reldebug configuration).
This is the dataset I used: https://rgw.cs.uwaterloo.ca/pyserini/data/msmarco-passage-openai-ada2.tar (~100GiB compressed), however the script below only uses the first 1M rows to save on time/RAM.
Below is a script that reliably reproduces the issue on the machine I use: