sqlite in some cases dumps views in wrong order. how to reproduce on linux:
1) Make the test database with 2 "view". v1 depends on v2. Recreate v2 cat <<_EOF_ | sqlite3 test.sqlite PRAGMA foreign_keys=OFF; BEGIN TRANSACTION; CREATE TABLE test ( "id1" integer(8,0), "id2" integer(8,0), "id3" integer(8,0) ); CREATE VIEW v2 as select id1 from test where id1; CREATE VIEW v1 as select t1.id2 from test as t1 LEFT JOIN v2 ON t1.id2=v2.id1; DROP VIEW v2; CREATE VIEW v2 as select id1 from test where id1; COMMIT; _EOF_ 2) Dump the base sqlite3 test.sqlite .dump > test.sqlite.dump cat test.sqlite.dump => PRAGMA foreign_keys=OFF; BEGIN TRANSACTION; CREATE TABLE test ( "id1" integer(8,0), "id2" integer(8,0), "id3" integer(8,0) ); CREATE VIEW v1 as select t1.id2 from test as t1 LEFT JOIN v2 ON t1.id2=v2.id1; CREATE VIEW v2 as select id1 from test where id1; COMMIT; Here "CREATE VIEW v1" and "CREATE VIEW v2" placed in wrong order. 3) try to restore cat test.sqlite.dump | sqlite3 test2.sqlite => Error: near line 4: no such table: main.v2 Is it the user responsibility to follow all dependencies and recreate "VIEW" tree ? Best Regards Tyumentsev Alexander _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
