search

trinodb / trino

Official repository of Trino, the distributed SQL query engine for big data, formerly known as PrestoSQL (https://trino.io)
https://trino.io
Apache License 2.0
9.85k stars 2.85k forks source link

Improve Broadcast join efficiency for FTE #16262

Open arhimondr opened 1 year ago

arhimondr commented 1 year ago

Try to avoid re-creating build side hash map for every task

losipiuk commented 1 year ago

It would be nice to determine in the first step how much time are we wasting on building hashmap multiple times over whole run of tpcds benchmark suite.

arhimondr commented 1 year ago

In TPC-DS broadcast joins are rather small (in kilobytes), however it is theoretically possible for broadcast tables to be in gigabytes range