Horton

Re: remaining sql/json patches

jian he <jian.universality@gmail.com>

View thread
From: jian he <jian.universality@gmail.com>
To: Amit Langote <amitlangote09@gmail.com>
Cc: Andres Freund <andres@anarazel.de>, Alvaro Herrera <alvherre@alvh.no-ip.org>, Erik Rijkers <er@xs4all.nl>, Andrew Dunstan <andrew@dunslane.net>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2024-01-06T00:44:00Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. SQL/JSON: Various improvements to SQL/JSON query function docs

  2. SQL/JSON: Fix some obsolete comments.

  3. SQL/JSON: Fix issues with DEFAULT .. ON ERROR / EMPTY

  4. JSON_TABLE: Add support for NESTED paths and columns

  5. Fix JsonExpr deparsing to emit QUOTES and WRAPPER correctly

  6. Fix typo introduced in 6185c9737

  7. Add basic JSON_TABLE() functionality

  8. Avoid splitting errmsg string to span multiple lines

  9. Add SQL/JSON query functions

  10. Implement various jsonpath methods

  11. Add soft error handling to some expression nodes

  12. Adjust populate_record_field() to handle errors softly

  13. Refactor code used by jsonpath executor to fetch variables

  14. Test EXPLAIN (FORMAT JSON) ... XMLTABLE

  15. Simplify productions for FORMAT JSON [ ENCODING name ]

  16. Add trailing commas to enum definitions

  17. doc: add missing <returnvalue> and whitespace

  18. Add more SQL/JSON constructor functions

  19. Rename a nonterminal used in SQL/JSON grammar

  20. Some refactoring to export json(b) conversion functions

  21. Don't include CaseTestExpr in JsonValueExpr.formatted_expr

  22. Code review for commit b6e1157e7d

  23. Pass constructName to transformJsonValueExpr()

  24. Unify JSON categorize type API and export for external use

  25. Make some indentation in gram.y consistent

  26. Allow most keywords to be used as column labels without requiring AS.

  27. Reduce size of backend scanner's tables.

  28. Use perfect hashing, instead of binary search, for keyword lookup.

some tests after applying V33 and my small changes.
setup:
create table test_scalar1(js jsonb);
insert into test_scalar1 select jsonb '{"a":"[12,13]"}' FROM
generate_series(1,1e5) g;
create table test_scalar2(js jsonb);
insert into test_scalar2 select jsonb '{"a":12}' FROM generate_series(1,1e5) g;
create table test_array1(js jsonb);
insert into test_array1 select jsonb '{"a":[1,2,3,4,5]}' FROM
generate_series(1,1e5) g;
create table test_array2(js jsonb);
insert into test_array2 select jsonb '{"a": "{1,2,3,4,5}"}' FROM
generate_series(1,1e5) g;

tests:
----------------------------------------return a scalar int4range
explain(costs off,analyze) SELECT item FROM test_scalar1,
JSON_TABLE(js, '$.a' COLUMNS (item int4range PATH '$' omit quotes))
\watch count=5
237.753 ms

explain(costs off,analyze) select json_query(js, '$.a' returning
int4range omit quotes) from test_scalar1  \watch count=5
462.379 ms

explain(costs off,analyze) select json_value(js,'$.a' returning
int4range) from test_scalar1 \watch count=5
362.148 ms

explain(costs off,analyze) select (js->>'a')::int4range from
test_scalar1 \watch count=5
301.089 ms

explain(costs off,analyze) select trim(both '"' from
jsonb_path_query_first(js,'$.a')::text)::int4range from test_scalar1
\watch count=5
643.337 ms

----------------------------return a numeric array from jsonb array.
explain(costs off,analyze) SELECT item FROM test_array1,
JSON_TABLE(js, '$.a' COLUMNS (item numeric[] PATH '$')) \watch count=5
727.807 ms

explain(costs off,analyze) SELECT json_query(js, '$.a' returning
numeric[]) from test_array1 \watch count=5
2995.909 ms

explain(costs off,analyze) SELECT
replace(replace(js->>'a','[','{'),']','}')::numeric[] from test_array1
\watch count=5
2990.114 ms

----------------------------return a numeric array from jsonb string
explain(costs off,analyze) SELECT item FROM test_array2,
JSON_TABLE(js, '$.a' COLUMNS (item numeric[] PATH '$' omit quotes))
\watch count=5
237.863 ms

explain(costs off,analyze) SELECT json_query(js,'$.a' returning
numeric[] omit quotes) from test_array2 \watch count=5
893.888 ms

explain(costs off,analyze) SELECT trim(both '"'
from(jsonb_path_query(js,'$.a')::text))::numeric[] from test_array2
\watch count=5
1329.713 ms

explain(costs off,analyze) SELECT (js->>'a')::numeric[] from
test_array2 \watch count=5
740.645 ms

explain(costs off,analyze) SELECT trim(both '"' from
(json_query(js,'$.a' returning text)))::numeric[]  from test_array2
\watch count=5
1085.230 ms
----------------------------return a scalar numeric
explain(costs off,analyze) SELECT item FROM test_scalar2,
JSON_TABLE(js, '$.a' COLUMNS (item numeric PATH '$' omit quotes)) \watch count=5
238.036 ms

explain(costs off,analyze) select json_query(js,'$.a' returning
numeric) from test_scalar2 \watch count=5
300.862 ms

explain(costs off,analyze) select json_value(js,'$.a' returning
numeric) from test_scalar2 \watch count=5
160.035 ms

explain(costs off,analyze) select
jsonb_path_query_first(js,'$.a')::numeric from test_scalar2 \watch
count=5
294.666 ms

explain(costs off,analyze) select jsonb_path_query(js,'$.a')::numeric
from test_scalar2 \watch count=5
547.130 ms

explain(costs off,analyze) select (js->>'a')::numeric from
test_scalar2 \watch count=5
243.652 ms

explain(costs off,analyze) select (js->>'a')::numeric,
(js->>'a')::numeric from test_scalar2 \watch count=5
403.183 ms

explain(costs off,analyze) select json_value(js,'$.a' returning numeric),
        json_value(js,'$.a' returning numeric) from test_scalar2 \watch count=5
246.405 ms

explain(costs off,analyze) select json_query(js,'$.a' returning numeric),
        json_query(js,'$.a' returning numeric) from test_scalar2 \watch count=5
520.754 ms

explain(costs off,analyze) SELECT item, item1 FROM test_scalar2,
JSON_TABLE(js, '$.a' COLUMNS (item numeric PATH '$' omit quotes,
                            item1 numeric PATH '$' omit quotes)) \watch count=5
242.586 ms
---------------------------------
overall, json_value is faster than json_query. but json_value can not
deal with arrays in some cases.
but as you can see, in some cases, json_value and json_query are not
as fast as our current implementation.

Here I only test simple nested levels. if you extra multiple values
from jsonb to sql type, then json_table is faster.
In almost all cases, json_table is faster.

json_table is actually called json_value_op, json_query_op under the hood.
Without json_value and json_query related code, json_table cannot be
implemented.