Nested hash joins
Dear Team,
We are on 19.28 version. Can you please help us understand if there will be any issues with repeated hash joins as in below plan.
In a HASH JOIN, when a table is used as the first child (build input), does Oracle retain the entire dataset in memory until the probe side, especially when it is another nested HASH JOIN, starts producing rows?
===========================================
| Id | Operation | Name | Rows | Cost | Time | Start | Execs | Rows | Write | Write | Mem | Temp | Activity | Activity Detail |
| | | | (Estim) | | Active(s) | Active | | (Actual) | Reqs | Bytes | | | (%) | (# samples) |
=========================================================================================================================================================================
| -> 60 | HASH JOIN | | 101 | 9450 | 561 | +1 | 1 | 0 | 527K | 60GB | 294MB | 64GB | | |
| -> 61 | HASH JOIN | | 801 | 8836 | 559 | +3 | 1 | 530M | | | 8MB | . | | |
Tagged:
0