On 03/09/2017 06:28 AM, Vermes Mátyás wrote:
Why are you wanting me to work so hard at this?
Under no circumstances work hard, I rewrote the program for you to C. The
results are attached also, so you do not have to compile it, if you do not
want. You can download it from http://comfirm.hu/pub/sqlite3-regression.tar.gz.
You're updating the "megnevezes" field from within a loop on this SELECT:
select szamla, megnevezes from konto.proba
With 3.11.0, the scan is probably using the index instead of the table
(since it is smaller on disk and therefore faster):
create index konto.proba_nev on proba (megnevezes,szamla)
So the update is modifying the index while you are iterating through it.
The results of that can be unpredictable. If you change the SELECT to:
select szamla, megnevezes from konto.proba ORDER BY rowid
the SELECT will use the main table instead of the index and things will
likely work again.
Dan.
_______________________________________________
sqlite-users mailing list
sqlite-users@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
