On 16.05.2018 12:32, Richard Hipp wrote:
> On 5/16/18, Torsten Landschoff <torsten.landsch...@scale.eu> wrote:
>>
>> Actually it is: SQLite will generate the same query plan for both
>> queries as long as no statistics data is available. Adding the
>> statistics information will lead to the more complex and slow query plan
>> which is why I kept the statistics in the example SQL.
>
> Can you please send the output of ".fullschema" so that I can try to
> reproduce your problem?
>
Here you go:
---
torsten.landschoff@horatio:~$ /opt/sqlite3/bin/sqlite3 demo.db
SQLite version 3.23.2 2018-05-09 02:23:29
Enter ".help" for usage hints.
sqlite> .fullschema
CREATE TABLE base (
id INTEGER NOT NULL PRIMARY KEY,
modification_time DATETIME NOT NULL
);
CREATE UNIQUE INDEX ix_base_mtime ON base (modification_time);
CREATE TABLE derived (id INTEGER NOT NULL PRIMARY KEY);
ANALYZE sqlite_master;
ANALYZE sqlite_master;
ANALYZE sqlite_master;
---
Please note that I included a full example to reproduce the problem just
by running
---
sqlite3 < demo.sql
---
It's in my own reply to the original email.
Thank you for looking at this!
Greetings, Torsten
--
$---+----1----+----2----+----3----+----4----+----5----+----6----+
SCALE GmbH
Niederlassung Dresden
Torsten Landschoff
Pohlandstraße 19
01309 Dresden
Tel: +49-351-312002-10
Fax: +49-351-312002-29
SCALE GmbH
Registergericht und Sitz: Ingolstadt, HRB 6384
Geschäftsführer: Dr.-Ing. Heiner Müllerschön, Dipl.-Math. Ulrich Franz
_______________________________________________
sqlite-users mailing list
sqlite-users@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
