Is your feature request related to a problem? Please describe.
In a hash join, such lines will appear. I don't see them in the graphical plan, although they may help to estimate how high work_mem×hash_mem_mutliplier must be to avoid temp files:
Buckets: 1048576 Batches: 16 Memory Usage: 62533kB
Buffers: shared hit=14909 read=98728, temp written=88109
(Source, JSON version )
Another example with a HashAggregate here, with a another field, Planned Partitions:
Describe the solution you'd like
- Print somewhere the information about Buckets and Batches and Memory, and the total ( about 1 GB in the example)
Describe alternatives you've considered
Temporary files information is not always accurate enough (here only 690 MB, although the hash join wants 1 GB to run in memory)
Additional context
I have a script in a private repo, and it will probably be included in dali.bo/j6_html soon.
This information was already there in PG 9.1.
Is your feature request related to a problem? Please describe.
In a hash join, such lines will appear. I don't see them in the graphical plan, although they may help to estimate how high
work_mem×hash_mem_mutlipliermust be to avoid temp files:(Source, JSON version )
Another example with a HashAggregate here, with a another field,
Planned Partitions:Describe the solution you'd like
Describe alternatives you've considered
Temporary files information is not always accurate enough (here only 690 MB, although the hash join wants 1 GB to run in memory)
Additional context
I have a script in a private repo, and it will probably be included in dali.bo/j6_html soon.
This information was already there in PG 9.1.