Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-05-02 Thread Mario M. Westphal
Thank you for providing this pre-release amalgamation ;-) I downloaded it immediately and compiled it into my software. The problem has been ++resolved++ and the performance is at least as good as with previous versions of SQLite. It even feels a bit faster, although I only could try it with a 20

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-30 Thread Mario M. Westphal
Thank you for providing this pre-release amalgamation ;-) I downloaded it immediately and compiled it into my software. The problem has been ++resolved++ and the performance is at least as good as with previous versions of SQLite. It even feels a bit faster, although I only could try it with a 20

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-30 Thread Richard Hipp
On Tue, Apr 30, 2013 at 3:34 AM, Mario M. Westphal wrote: > Hi, > > I just tried this (sorry dor the delay) but apparently I'm not having the > right toolset installed. > Please download the latest amalgamations from http://www.sqlite.org/draft/download.html -- D. Richard Hipp d...@sqlite.org

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-30 Thread Mario M. Westphal
Hi, I just tried this (sorry dor the delay) but apparently I'm not having the right toolset installed. The make file requires gawk.exe (which I downloaded from sourceforge), but now it's complaing about a missing tclsh85... Since the previous SQLite version works fine I think I'll skip this and w

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-22 Thread Richard Hipp
On Mon, Apr 22, 2013 at 8:20 AM, Mario M. Westphal wrote: > > So far I only used the Amalgamation. Looks like the files on your source > control server require me to build SQLite from source or at least run a > tool/script to build the amalgamation. > > I'm using Windows and Visual Studio so the

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-22 Thread Mario M. Westphal
So far I only used the Amalgamation. Looks like the files on your source control server require me to build SQLite from source or at least run a tool/script to build the amalgamation. I'm using Windows and Visual Studio so the standard Linux build tools and scripting languages are not available.

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-21 Thread Richard Hipp
On Thu, Apr 18, 2013 at 3:27 AM, Mario M. Westphal wrote: > > After downloading and compiling in the SQLite 3.7.16.2, we have a problem. > > The INSERT statement which takes only a few seconds with previous builds > now does not return, at least not within several minutes. > It's hard to debug th

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-19 Thread Richard Hipp
FWIW: The following script demonstrates the problem: CREATE TABLE t1(x INTEGER, y INTEGER); CREATE INDEX t1xy ON t1(x,y); CREATE INDEX t1y ON t1(y); CREATE TABLE t2(x INTEGER PRIMARY KEY, y INTEGER); CREATE TABLE t3(x INTEGER PRIMARY KEY, y INTEGER); CREATE INDEX t3y ON t3(y); .wheretrace off

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-19 Thread Mario M. Westphal
I'm currently uploading the database and will send you a link to it. -- Mario ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-19 Thread Richard Hipp
On Fri, Apr 19, 2013 at 9:17 AM, Mario M. Westphal wrote: > Hi, thanks. > > But I'm not sure that I understand you correctly. > > Is this behavior considered as something that needs to be looked at by the > SQLite Team to restore the original performance, or is this how "it is" now > and I have t

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-19 Thread Mario M. Westphal
Hi, thanks. But I'm not sure that I understand you correctly. Is this behavior considered as something that needs to be looked at by the SQLite Team to restore the original performance, or is this how "it is" now and I have to find a work-around for good (e.g. applying your suggestion with CROS

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-19 Thread Richard Hipp
On Thu, Apr 18, 2013 at 3:27 AM, Mario M. Westphal wrote: > We can provide a sample database etc. on request. > Can you send the database to my private email, please? -- D. Richard Hipp d...@sqlite.org ___ sqlite-users mailing list sqlite-users@sqlit

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-19 Thread Richard Hipp
The change in behavior occurs with http://www.sqlite.org/src/info/38852f158a If you need a work-around, change INNER to CROSS in the queries and the old query plan will be restored. On Thu, Apr 18, 2013 at 3:27 AM, Mario M. Westphal wrote: > This is a SQL Script > > > /* > Application linking t

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-19 Thread support
Hi, Celemens I created the requested log data by running the same operations in my application, once compiled with the 3.7.15.1 version (OK) and once with the 3.7.16.2 (BAD performance). The query plans are quite different. 1. --- The query my application runs for this test INSERT INTO md_fts_c

Re: [sqlite] Potential problem in SQLite 3.7.16.2

2013-04-18 Thread Clemens Ladisch
Mario M. Westphal wrote: > The INSERT statement which takes only a few seconds with previous builds > now does not return, at least not within several minutes. Please show the output of EXPLAIN QUERY PLAN for the SELECT part of the query in both versions. Regards, Clemens ___

[sqlite] Potential problem in SQLite 3.7.16.2

2013-04-18 Thread Mario M. Westphal
This is a SQL Script /* Application linking to SQLite using the Amalgation. Build Tool: C++, Visual Studio 2012, Windows 7 64-Bit The table schema and the query enclosed below are in use for over one year and various SQLite versions. After downloading and compiling in the SQLite 3.7.16.2, we h