On Thu, Oct 17, 2024 at 09:21:11AM +0200, Anthonin Bonnefoy wrote: > I'm not sure about this approach. It moves the responsibility of > tracking the location and length from the nested statement to the top > level statement. > - The location you added in ExplainStmt and CopyStmt has a different > meaning from all others and tracks the nested location and not the > location of the statement itself. This creates some inconsistencies. > - The work will need to be done for all statements with nested > queries: Prepare, Create table as, Declare Cursor, Materialised View. > Whereas by tracking the location of PreparableStatements, there's no > need for additional logic. For example, v8 0002 fixes the existing > behaviour for Prepare statements thanks to SelectStmt's modifications.
Hmm. And isn't tracking this information at only the top-level going to be a problem when dealing with multiple levels of nesting, when for example these involve pl/sql or pl/pgSQL functions? How would this work if we're dealing with multiple levels of Nodes? -- Michael
signature.asc
Description: PGP signature
