pg_query could not detect length of RawStmt(s) with quoted characters

[valerysvl]

Example query with 2 stataments:

UPDATE _ScheduledJobs23816 SET _RestartAttemptNumber = CAST(0 AS NUMERIC), _State = CAST(1 AS NUMERIC), _StartTime = '2023-05-16 11:22:52'::timestamp, _FinishTime = '2023-05-16 11:21:27'::timestamp WHERE _ScheduledJobs23816._ID = '\244\331\340\376\225\250\343\343A\005Z3\\270c\\273'::bytea; UPDATE _ScheduledJobs23816 SET _RestartAttemptNumber = CAST(0 AS NUMERIC), _State = CAST(1 AS NUMERIC), _StartTime = '2023-05-16 11:22:52'::timestamp, _FinishTime = '2023-05-16 11:21:27'::timestamp WHERE _ScheduledJobs23816._ID = '\\244\\331\\340\\376\\225\\250\\343\\343A\\005Z3\270c\273'::bytea;

pg_query can not detect on RawStmt list:

1. correct stmt_length of both statements,
2. correct stmt_location on 2nd statement

Used pg_query for PG15.1

[lfittl]

@valerysvl Thanks for the report!

Could you clarify what you are expecting, and what you are seeing?

From a quick test using pg_query, this appears to work correctly:

irb(main):021:0> puts x
UPDATE _ScheduledJobs23816 SET _RestartAttemptNumber = CAST(0 AS NUMERIC), _State = CAST(1 AS NUMERIC), _StartTime = '2023-05-16 11:22:52'::timestamp, _FinishTime = '2023-05-16 11:21:27'::timestamp
WHERE _ScheduledJobs23816._ID = '\244\331\340\376\225\250\343\343A\005`Z3\270c\273'::bytea;
UPDATE _ScheduledJobs23816 SET _RestartAttemptNumber = CAST(0 AS NUMERIC), _State = CAST(1 AS NUMERIC), _StartTime = '2023-05-16 11:22:52'::timestamp, _FinishTime = '2023-05-16 11:21:27'::timestamp
WHERE _ScheduledJobs23816._ID = '\244\331\340\376\225\250\343\343A\005`Z3\270c\273'::bytea;

irb(main):015:0> PgQuery.parse(x).tree.stmts.map {|s| [s.stmt_location, s.stmt_len] }
=> [[0, 288], [289, 289]]

irb(main):018:0> x[..288]
=> "UPDATE _ScheduledJobs23816 SET _RestartAttemptNumber = CAST(0 AS NUMERIC), _State = CAST(1 AS NUMERIC), _StartTime = '2023-05-16 11:22:52'::timestamp, _FinishTime = '2023-05-16 11:21:27'::timestamp\nWHERE _ScheduledJobs23816._ID = '\\244\\331\\340\\376\\225\\250\\343\\343A\\005`Z3\\270c\\273'::bytea;"

irb(main):019:0> x[289..][..289]
=> "\nUPDATE _ScheduledJobs23816 SET _RestartAttemptNumber = CAST(0 AS NUMERIC), _State = CAST(1 AS NUMERIC), _StartTime = '2023-05-16 11:22:52'::timestamp, _FinishTime = '2023-05-16 11:21:27'::timestamp\nWHERE _ScheduledJobs23816._ID = '\\244\\331\\340\\376\\225\\250\\343\\343A\\005`Z3\\270c\\273'::bytea;"

[valerysvl]

after parsing protobuf_result = pg_query_parse_protobuf(source_query);

and following unpacking unpacked = pg_query__parse_result__unpack(NULL, pgquery_pbparse_result.parse_tree.len, pgquery_pbparse_result.parse_tree.data)

I am expecting stms_location as correct offset on source_query

unpacked.stmts[0].stmt_len = 289; unpacked.stmts[0].stmt_location = 0; unpacked.stmts[1].stmt_len = 289; unpacked.stmts[1].stmt_location = 289; unpacked.stmts[2].stmt_len = 289; unpacked.stmts[2].stmt_location = 579; unpacked.stmts[3].stmt_len = 289; unpacked.stmts[3].stmt_location = 867; ...

[lfittl]

@valerysvl Its not clear to me what you mean with "correct offset on source_query" - from the data provided the len and location appear to be correct.

Could you provide a fully self-contained example as a reproducer?