> On Fri, Aug 4, 2017 at 7:41 AM, petern <[email protected]>
>> wrote:
>>
>>> Sylvain, are you happy with the performance?
>>>
>>> Maybe you are using it differently but, from my tests, the DEFAULT clause
>>> is ignored for PRIMARY KEY columns. I had to use an ordinary column
>>> with
>>> UNIQUE constraint to test your extension. Below is a tester for 1
>>> million
>>> rows which completes in about 186 seconds. The same million row test
>>> with
>>> PRIMARY KEY column (and ignored DEFAULT) completes in about 5 seconds.
>>>
>>
> I am sorry I fail to see how to improve it, I would be really interested
> to have advises from the experts!
>
HURRAAAA ! I found something !
if I remove the checks on the SP_SEQUENCE table, it is much faster
sqlite> WITH T(i) AS (SELECT (1)i UNION ALL SELECT i+1 FROM T WHERE
i<1000000) INSERT INTO seq_test(seq_num) SELECT seq_nextval('seq1') from T;
Run Time: real 18.735 user 16.614106 sys 0.109201
and I have another idea to select on the rowid at the update.
Best regards,
Sylvain
_______________________________________________
sqlite-users mailing list
[email protected]
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
