2011/8/7 Simon Slavin <[email protected]>:
> You don't need to. The SQLite expressions I listed tell you how to achieve
> the result without doing that.
Really? And how can you perform the query like to:
sqlite> create table t1(ids text);
sqlite> insert into t1 (ids) values ('1 2 3');
sqlite> insert into t1 (ids) values ('2 3 4');
sqlite> insert into t1 (ids) values ('3 4 5');
sqlite> create table t2(name text);
sqlite> insert into t2 (name) values ('name1');
sqlite> insert into t2 (name) values ('name2');
sqlite> insert into t2 (name) values ('name3');
sqlite> insert into t2 (name) values ('name4');
sqlite> insert into t2 (name) values ('name5');
sqlite> select * from t2 where rowid in (select ids from t1 where rowid=2);
A simple calculation: if each list of identifiers have about 1000
items and there are
1 000 000 lists than the table of relations (t1.rowid, t2.rowid) will
have 1 000 000 000
rows! It's too slow and is not useful in real world. Of cource all
systems store lists of
identifiers in similar situations.
--
Best regards, Alexey Pechnikov.
http://pechnikov.tel/
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users