Thanks. > the corruption is restricted to indices, so that it can be completely > repaired by running "REINDEX"
Yes, REINDEX helps at least to the extent that integrity_check succeeds. Question: How can we recognize that integrity errors refer solely to an index corruption? (Considering the possibility of automatic REINDEX.) I refer to the index sqlite_autoindex_account_1, which I did not consider to be an ordinary index. > Though, the fact that the corruption is restricted to an index is probably > a coincidence. Probably not a coincidence as all cases where we have additional data point to an index problem. ------------ Meanwhile I knew a bit more: Rows with strange ROWID refer to accounts that were first added to the DB (first part of the sync process), but immediately afterwards (during sync cleanup) deleted from the DB. The deletion was done within a transaction and used relatively complex SQL command. Sync action is blocking - the user has no DB access during the whole process. The crash happened during the first user action following the sync, which in this case was displaying the account list in a form. Best regards, Jan -- View this message in context: http://sqlite.1065341.n5.nabble.com/What-can-be-deduced-from-integrity-check-tp70451p70455.html Sent from the SQLite mailing list archive at Nabble.com. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users