Hi, Laura, On Wed, Sep 7, 2016 at 12:25 PM, Laura BERGOENS <laura.bergo...@imerir.com> wrote: > Hi Dominique, > > I guess it's not, but I'm not quite sure on what I can share, I'm a newbie > in a professional environment. I would be more confortable if I ask my boss > first. In the meantime, all I can say is that there are no primary keys nor > foreign keys in the table, and no indexes at all. I activated the > automatic_index PRAGMA. > > All tables look the same more or less : > > CREATE TABLE TABLEA (ID_A INTEGER,NUM_A INTEGER,NUM_GROUPE_A INTEGER,TYPE_A > VARCHAR(255),LIBELLE_A VARCHAR(255),NOM_A VARCHAR(255));
If you don't have keys nor indexes in any of the tables than the query time will definitely be slower. Now let me ask you this: Is there a reason why you can't create a copy of the database and work with a copy? Just open the Terminal and do: cp real_database.db my_database.db Then you can experiment with the database itself without screwing up the real data. Besides, that's how actual dev environment should be - you have access to the development database on which you do the experiments and then the DB Admin will populate the schema based on you analysis for production. Now, you need to create a primary/foreign keys and indexes on the tables you are creating. Then the execution time will improve even more. Try it and see. Then when you are ready ask the DB Admin to make those changes for production database. Thank you. BTW, if its so scary, you can make the field name as field1, field2, field3, etc. That way the design of the database and the application will stay as the company secret. Thank you. > > 2016-09-07 18:13 GMT+02:00 Dominique Devienne <ddevie...@gmail.com>: > >> On Wed, Sep 7, 2016 at 6:00 PM, Laura BERGOENS <laura.bergo...@imerir.com> >> wrote: >> >> > Then I create real tables in the in-memory clone that contain the content >> > of the views : INSERT INTO tableA SELECT * from viewA >> > >> >> What matters is how you create the tables, >> and in particular what primary key and indexes you use on them. >> >> If you don't have indexes (or a PK) on the columns you're joining on, >> it can't be as fast as it could be (and SQLite might end-up creating >> those "indexes" on-the-fly, and recreate them on the next query, etc...). >> >> So as advised, share your schema (see Richard's post), not your data, >> and share your queries and explain query plan for your queries. >> >> You're schema is not confidential, is it? The data, sure. But the schema??? >> --DD >> _______________________________________________ >> sqlite-users mailing list >> sqlite-users@mailinglists.sqlite.org >> http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users >> > > > > -- > Laura BERGOENS > Technicienne supérieure en Informatique et étudiante à l'IMERIR de Perpignan > > *Institut Méditerranéen d'Étude etde Recherche en Informatique* > _______________________________________________ > sqlite-users mailing list > sqlite-users@mailinglists.sqlite.org > http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
