Re: remaining sql/json patches

jian he <jian.universality@gmail.com>

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: 2023-12-25T05:03:27Z
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.

Hi.

+/*
+ * JsonTableFetchRow
+ * Prepare the next "current" tuple for upcoming GetValue calls.
+ * Returns FALSE if the row-filter expression returned no more rows.
+ */
+static bool
+JsonTableFetchRow(TableFuncScanState *state)
+{
+ JsonTableExecContext *cxt =
+ GetJsonTableExecContext(state, "JsonTableFetchRow");
+
+ if (cxt->empty)
+ return false;
+
+ return JsonTableScanNextRow(cxt->root);
+}

The declaration of struct JsonbTableRoutine, SetRowFilter field is
null. So I am confused by the above comment.
also seems the  `if (cxt->empty)` part never called.

+static inline JsonTableExecContext *
+GetJsonTableExecContext(TableFuncScanState *state, const char *fname)
+{
+ JsonTableExecContext *result;
+
+ if (!IsA(state, TableFuncScanState))
+ elog(ERROR, "%s called with invalid TableFuncScanState", fname);
+ result = (JsonTableExecContext *) state->opaque;
+ if (result->magic != JSON_TABLE_EXEC_CONTEXT_MAGIC)
+ elog(ERROR, "%s called with invalid TableFuncScanState", fname);
+
+ return result;
+}
I think Assert(IsA(state, TableFuncScanState)) would be better.

+/*
+ * JsonTablePlanType -
+ * flags for JSON_TABLE plan node types representation
+ */
+typedef enum JsonTablePlanType
+{
+ JSTP_DEFAULT,
+ JSTP_SIMPLE,
+ JSTP_JOINED,
+} JsonTablePlanType;
it would be better to add some comments on it. thanks.

JsonTablePlanNextRow is quite recursive! Adding more explanation would
be helpful, thanks.

+/* Recursively reset scan and its child nodes */
+static void
+JsonTableRescanRecursive(JsonTablePlanState * state)
+{
+ if (state->type == JSON_TABLE_JOIN_STATE)
+ {
+ JsonTableJoinState *join = (JsonTableJoinState *) state;
+
+ JsonTableRescanRecursive(join->left);
+ JsonTableRescanRecursive(join->right);
+ join->advanceRight = false;
+ }
+ else
+ {
+ JsonTableScanState *scan = (JsonTableScanState *) state;
+
+ Assert(state->type == JSON_TABLE_SCAN_STATE);
+ JsonTableRescan(scan);
+ if (scan->plan.nested)
+ JsonTableRescanRecursive(scan->plan.nested);
+ }
+}

From the coverage report, I noticed the first IF branch in
JsonTableRescanRecursive never called.

+ foreach(col, columns)
+ {
+ JsonTableColumn *rawc = castNode(JsonTableColumn, lfirst(col));
+ Oid typid;
+ int32 typmod;
+ Node   *colexpr;
+
+ if (rawc->name)
+ {
+ /* make sure column names are unique */
+ ListCell   *colname;
+
+ foreach(colname, tf->colnames)
+ if (!strcmp((const char *) colname, rawc->name))
+ ereport(ERROR,
+ (errcode(ERRCODE_SYNTAX_ERROR),
+ errmsg("column name \"%s\" is not unique",
+ rawc->name),
+ parser_errposition(pstate, rawc->location)));

this `/* make sure column names are unique */` logic part already
validated in isJsonTablePathNameDuplicate, so we don't need it?
actually isJsonTablePathNameDuplicate validates both column name and pathname.

select jt.* from jsonb_table_test jtt,
json_table (jtt.js,'strict $[*]' as p
columns (n for ordinality,
nested path 'strict $.b[*]' as pb columns ( c int path '$' ),
nested path 'strict $.b[*]' as pb columns ( s int path '$' ))
) jt;

ERROR:  duplicate JSON_TABLE column name: pb
HINT:  JSON_TABLE column names must be distinct from one another.
the error is not very accurate, since pb is a pathname?