cmu-db / benchbase

Multi-DBMS SQL Benchmarking Framework via JDBC
https://db.cs.cmu.edu/projects/benchbase/
Other
467 stars 184 forks source link

TPC-H schema contains many extra indexes #233

Open mariadb-DmitryVolkov opened 1 year ago

mariadb-DmitryVolkov commented 1 year ago

I've noticed that TPC-H schema contains many extra indexes (for quite some tables). Here is an example: line item table has at least 6 extra indexes:

https://github.com/cmu-db/benchbase/blob/main/src/main/resources/benchmarks/tpch/ddl-postgres.sql#L125-L130

Where are these indexes came from? I've tried to read the tpc-h description but couldn't find one. I must be blind today.

Looking at HammerDB, for example (this is just another popular tool, not a gold standard, of course) https://github.com/TPC-Council/HammerDB/blob/master/src/mariadb/mariaolap.tcl#L194-L215

it has some indexes, but not 6.

apavlo commented 1 year ago

@mariadb-DmitryVolkov Good point. We should look this over. All the different DBMS-specific DDL files have different indexes.