I think the point is that attempting to create the unique index should fail (with an error) and the index not be created rather than creating a unique index with duplicates (or whatever it is doing) causing subsequent queries to return incorrect results.
>On Wed, 30 Jul 2014 09:10:29 +0400 >Pavel Pimenov <pavel.pime...@gmail.com> wrote: > >> CREATE TABLE fly_hash_block(tth_id integer PRIMARY KEY NOT NULL, tth >> number NOT NULL); >> INSERT INTO fly_hash_block VALUES(1,1); >> INSERT INTO fly_hash_block VALUES(2,2); >> INSERT INTO fly_hash_block VALUES(3,2); >> *CREATE UNIQUE INDEX iu_fly_hash_block_tth ON fly_hash_block(tth);* > >Not an answer if the table already exists, but a better solution would >be to constrain the table: > >sqlite> CREATE TABLE fly_hash_block( > tth_id integer PRIMARY KEY NOT NULL, > tth number not null UNIQUE); >sqlite> INSERT INTO fly_hash_block VALUES(1,1); >sqlite> INSERT INTO fly_hash_block VALUES(2,2); >sqlite> INSERT INTO fly_hash_block VALUES(3,2); >Error: column tth is not unique >sqlite> > >--jkl >_______________________________________________ >sqlite-users mailing list >sqlite-users@sqlite.org >http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users