Hi,
While writing the recent ext.stats docs, I became annoyed by the output
functions of the new types used by multivariate statistics: they are
almost JSON, but not quite. Since they can become largish, I propose
that we make a point of ensuring the output of those types is valid
JSON, so that they can be processed by JSON tools, particularly by
jsonb_pretty(). Note that the internal bytea format does not change;
this is only for human consumption (or for external tools that want to
examine the values), not for the planner. Also, the input function for
those types rejects input, so nothing needs to care about reading these
values; ANALYZE is the only valid source for them.
With the proposed attached patch, those columns look more convenient to
work with:
ialvherre=# select jsonb_pretty(stxndistinct::jsonb), jsonb_pretty( stxdependencies::jsonb) from pg_statistic_ext;
jsonb_pretty │ jsonb_pretty
──────────────────────┼────────────────────────────
{ ↵│ { ↵
"1, 2": 33178, ↵│ "1 => 2": 1.000000, ↵
"1, 5": 33178, ↵│ "1 => 5": 1.000000, ↵
"2, 5": 27435, ↵│ "5 => 1": 0.422367, ↵
"1, 2, 5": 33178↵│ "5 => 2": 0.482567, ↵
} │ "1, 2 => 5": 1.000000,↵
│ "1, 5 => 2": 1.000000,↵
│ "2, 5 => 1": 0.807367 ↵
│ }
Changes:
* Removed the external [ ]; the whole string is now a single JSON object.
* Removed the bitmapset output artifact. (I introduced that while
simplifying the code, but I now think that was a mistake).
* The attribute list is the object key.
* In ndistinct, the value is now an integer rather than a floating point
number.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers