On 8/30/19 3:34 AM, Keith Medcalf wrote: > Fascinating again, because the code is identical. The p1 difference in the > OpenWrite/IdxInsert/IdxDelete opcodes is just the "file handle" that happens > to be being used and doesn't have any real significance (at least I don't > think it does). The p3 in the Transaction opcode is merely the schemacookie > and not only indicates that the schema has seen 2 changes between these plans > being generated (presumably the DROP and CREATE of the index). > > I am at a loss to explain the difference in execution speed.
Do not bother! I was just curious if this was expected behavior. Until now I put about 5h into solving this little riddle. With DROP INDEX / CREATE INDEX I got something that works for me and therefore I will leave it at that. However, if one wants to investigate further, I am happy to provide additional information. -- Best regards dirdi _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users