-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Wanadoo Hartwig wrote: >> Actually, if you replace the FTS3 table with any other type of table - >> or as many tables as you like and do also in these tables insert >> operations - the last row id of the original table is correctly >> returned. Therefore, I assume (I do not have any details) that there >> is a bug in the trigger mechanism itself.
The point that keeps being missed is that FTS3 is not a regular table. It is a virtual table. That means that the SQLite core calls it corresponding to various SQL operations but behind the scenes it is up the virtual table implementation what to do. Read about them at: http://www.sqlite.org/cvstrac/wiki?p=VirtualTables A virtual table implementation is free to do whatever it wants. When asked to insert one row, it could store 27 after converting all text to Klingon. It could instead delete whatever you asked it to add. It could do nothing. The FTS3 implementation happens to have 3 other tables and so on an insert those are updated with extra information and you are seeing those after effects. Roger -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkrk/A4ACgkQmOOfHg372QTEWACggOjPYsHFzB00jNMcDkOmYQ5q KCgAniNN8LAdKea5ZhHYO5SgrsG7qhdw =R9aD -----END PGP SIGNATURE----- _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
