Michal Nowakiewicz created ARROW-17204:
------------------------------------------
Summary: [C++][Compute] Switch new (Swiss) hash join to use 64-bit
hash
Key: ARROW-17204
URL: https://issues.apache.org/jira/browse/ARROW-17204
Project: Apache Arrow
Issue Type: Improvement
Components: C++
Affects Versions: 10.0.0
Reporter: Michal Nowakiewicz
Fix For: 10.0.0
Currently SwissJoin uses 32-bit hash. This means that it will work well only up
to around 10M rows in the hash table. Beyond that the hash collisions will keep
growing very fast making the hash join impractical to use from performance
perspective. Switch to 64-bit hash.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
