Re: ANALYZE ONLY - Mailing list pgsql-hackers

From Melih Mutlu
Subject Re: ANALYZE ONLY
Date
Msg-id CAGPVpCT2_fCP-GgObXNOkMXEvKU2L6Gud2ABpGfRQ8--9ed=WA@mail.gmail.com
Whole thread Raw
In response to ANALYZE ONLY  (Michael Harris <harmic@gmail.com>)
List pgsql-hackers
Hi Michael,

Thanks for the patch.
I quickly tried running some ANALYZE ONLY queries, it seems like it works fine.

-ANALYZE [ ( <replaceable class="parameter">option</replaceable> [, ...] ) ] [ <replaceable class="parameter">table_and_columns</replaceable> [, ...] ]
+ANALYZE [ ( <replaceable class="parameter">option</replaceable> [, ...] ) ] [ [ ONLY ] <replaceable class="parameter">table_and_columns</replaceable> [, ...] ]

It seems like extended_relation_expr allows "tablename *" syntax too. That should be added in docs as well. (Same for VACUUM doc)

 <para>
    For partitioned tables, <command>ANALYZE</command> gathers statistics by
    sampling rows from all partitions; in addition, it will recurse into each
    partition and update its statistics.  Each leaf partition is analyzed only
    once, even with multi-level partitioning.  No statistics are collected for
    only the parent table (without data from its partitions), because with
    partitioning it's guaranteed to be empty.
  </para>

We may also want to update the above note in ANALYZE doc.

+-- ANALYZE ONLY / VACUUM ONLY on partitioned table
+CREATE TABLE only_parted (a int, b char) PARTITION BY LIST (a);
+CREATE TABLE only_parted1 PARTITION OF vacparted FOR VALUES IN (1);
+INSERT INTO only_parted1 VALUES (1, 'a');

Tests don't seem right to me.
I believe the above should be " PARTITION OF only_parted " instead of vacparted.
It may better to insert into partitioned table (only_parted) instead of the partition (only_parted1);

Also it may be a good idea to test VACUUM ONLY for inheritance tables the same way you test ANALYZE ONLY.

Lastly, the patch includes an unrelated file (compile_flags.txt) and has whitespace errors when I apply it.

Regards,
--
Melih Mutlu
Microsoft

pgsql-hackers by date:

Previous
From: David Rowley
Date:
Subject: Re: Redundant Result node
Next
From: Rafia Sabih
Date:
Subject: Re: Support tid range scan in parallel?