Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-28 Thread Bart Smissaert
> I for my part would *insist* on SQLite throwing an error, when it encounters unknown (userdefined) functions or collations Yes, sure, you need to know about the missing UDF or collation. I suppose there is no sensible way to know about this and still be able to run the SQL. I looked at the DateP

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-28 Thread Olaf Schmidt
Am 28.10.2017 um 23:00 schrieb Richard Hipp: On 10/27/17, Olaf Schmidt wrote: The new CoRoutine-approach seems to slow down certain ViewDefinitions (in comparison to running a Query directly). Can you please download and try the latest "Prerelease Snapshot" from https://sqlite.org/download.ht

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-28 Thread Olaf Schmidt
Am 28.10.2017 um 19:22 schrieb Bart Smissaert: This is interesting, not so much for the performance regression, but for the fact that the posted Northwind database has spaces in table names and view names. I wasn't aware this is allowed and it caused a lot of errors in my app. I have this nearly

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-28 Thread Richard Hipp
On 10/27/17, Olaf Schmidt wrote: > The new CoRoutine-approach seems to slow down certain > ViewDefinitions (in comparison to running a Query directly). Can you please download and try the latest "Prerelease Snapshot" from https://sqlite.org/download.html and let me know whether or not it clears y

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-28 Thread Bart Smissaert
This is interesting, not so much for the performance regression, but for the fact that the posted Northwind database has spaces in table names and view names. I wasn't aware this is allowed and it caused a lot of errors in my app. I have this nearly fixed now (by adding the square brackets), but I

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-27 Thread Olaf Schmidt
Am 27.10.2017 um 21:11 schrieb Richard Hipp: Thanks for the report. Do you have any other interesting, complex, or slow queries using your database that you can send me for testing purposes? With regards to the NWind.db, the Invoices-View is the one with the largest processing-time. NWind.db

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-27 Thread Olaf Schmidt
Am 27.10.2017 um 21:59 schrieb David Raymond: Also getting 31ms for both... Thanks for testing that guys... I was able to get similar timings (about 30msec), when the (quite large) resultset-output was delegated into a file... To avoid doing expensive File-IO in that test, I'd recommend to run

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-27 Thread David Raymond
PM To: SQLite mailing list Subject: Re: [sqlite] Small Performance Regression (3.21.0) I don't see any difference in the runtimes, at least not with the current head of trunk ... sqlite> .once x sqlite> select * from invoices; Run Time: real 0.032 user 0.031250 sys 0.00 sqlite

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-27 Thread Richard Hipp
Thanks for the report. Do you have any other interesting, complex, or slow queries using your database that you can send me for testing purposes? On 10/27/17, Olaf Schmidt wrote: > The new CoRoutine-approach seems to slow down certain > ViewDefinitions (in comparison to running a Query directly)

Re: [sqlite] Small Performance Regression (3.21.0)

2017-10-27 Thread Keith Medcalf
ay to Heaven says a lot about anticipated traffic volume. >-Original Message- >From: sqlite-users [mailto:sqlite-users- >boun...@mailinglists.sqlite.org] On Behalf Of Olaf Schmidt >Sent: Friday, 27 October, 2017 12:21 >To: sqlite-users@mailinglists.sqlite.org >Subject: [s

[sqlite] Small Performance Regression (3.21.0)

2017-10-27 Thread Olaf Schmidt
The new CoRoutine-approach seems to slow down certain ViewDefinitions (in comparison to running a Query directly). FWIW, here's a download-link to an NorthWind-SQLite-DB, which already contains certain view-definitions: http://vbRichClient.com/Downloads/NWind.zip (an "Analyze"-command was already