Closed wendigo closed 1 year ago
For even a simple query performance is much much better:
After:
trino> select lower(orderstatus), count(1) from tpch.sf100.orders group by lower(orderstatus); _col0 | _col1 -------+---------- p | 3841445 o | 73086053 f | 73072502 (3 rows) Query 20230413_162027_00025_v32u9, FINISHED, 3 nodes Splits: 54 total, 54 done (100.00%) 9.38 [150M rows, 0B] [16M rows/s, 0B/s]
Before:
trino> select lower(orderstatus), count(1) from tpch.sf100.orders group by lower(orderstatus); _col0 | _col1 -------+---------- p | 3841445 o | 73086053 f | 73072502 (3 rows) Query 20230413_171901_00025_x7epn, FINISHED, 3 nodes Splits: 54 total, 54 done (100.00%) 15.32 [150M rows, 0B] [9.79M rows/s, 0B/s]
For even a simple query performance is much much better:
After:
Before: