*** grantc has joined #ingres | 00:13 | |
*** grantc has quit IRC | 00:13 | |
*** grantc has joined #ingres | 00:14 | |
*** KermitTheFragger has joined #ingres | 00:34 | |
*** atrofast has joined #ingres | 02:32 | |
*** atrofast has quit IRC | 02:37 | |
*** atrofast has joined #ingres | 02:38 | |
*** grantc has quit IRC | 02:58 | |
*** grantc has joined #ingres | 03:14 | |
pboro | atrofast, geotools stuff progressing well? :) | 03:41 |
---|---|---|
atrofast | pboro: pretty well, there seems to be a bug in MVCC and LOB handling that needs resolving | 03:48 |
pboro | nice to hear | 03:50 |
atrofast | Yeah... oh how familiar are you with JDBC and LOBs anyway pboro? Can you actually insert a LONG VARCHAR longer than 32000 bytes NOT using PreparedStatements? | 04:00 |
pboro | as far as I know, you can't | 04:00 |
atrofast | So you gotta use PreparedStatements for LOBs the neh? | 04:01 |
pboro | yup | 04:01 |
pboro | I usually only use prepared statements for the security they provided by using placeholders | 04:02 |
atrofast | Good to know... I wrote a second driver for GeoTools (utilizing much of the first one) last night but to use Prepared Statements instead... was pretty sure something like that would come up :) | 04:02 |
atrofast | How does performance compare? | 04:02 |
atrofast | prepared vs not prepared | 04:02 |
pboro | well the server needs to "prepare" the statement any way, you just get a possibility to reuse the once already prepared statement...if you're using statement pooling, the jdbc pooling library can keep the statements in a "pooled cache" | 04:03 |
pboro | so you should get benefits in that case, since the statement only needs to be prepared once for each connection | 04:04 |
pboro | pooling libraries like c3p0 and apache dbcp support also pooling the statements, which is nice | 04:05 |
pboro | but of course there's a small penalty in using preparedstatement since the statement will be "stored" for a while etc., but if you really need performance you will need to use prepared statement to make them cacheable/poolable | 04:05 |
pboro | I might be wrong, but this is how I understand and do it :) | 04:06 |
pboro | of course you can also reuse the statements in your code, without a caching/pooling library between your application and jdbc driver | 04:07 |
atrofast | That makes sense... I'm not sure what kind of strategy GeoTools uses if it's pooling... But I guess it's a moot point since I need to prepare my statements in case the geometries are large | 04:09 |
pboro | mostly the pooling library is put between the application and the jdbc driver and the software doesn't take any part in it | 04:10 |
pboro | the library just works as a caching layer | 04:10 |
pboro | you can use it or leave it out, either way you like | 04:10 |
pboro | in some cases it's a must, like some high volume web applications | 04:11 |
pboro | all java web application frameworks only use prepared statements, like Spring Framework | 04:11 |
pboro | so... I would really go with the prepared ones :) | 04:11 |
atrofast | Thanks for the input, very useful :) Yeah looking over the other drivers, Oracle has ONLY a PreparedStatement driver. Postgres has both... MySQL and SQLServer has no Prepared Statement driver | 04:15 |
pboro | prepared statements is something that MySQL lacked a long time even for PHP | 04:15 |
pboro | MySQL 5.1+ does support prepared statements in it's JDBC driver | 04:16 |
pboro | maybe older too, can't find good source on the versions right now... | 04:17 |
pboro | support for prepared statements on the server side was added in 4.1 | 04:17 |
pboro | hehe, some statistics: http://euedge.com/blog/2007/11/11/prepared-statement-performance-in-mysql/ | 04:18 |
pboro | not too good test tho' | 04:19 |
*** javahorn has joined #ingres | 04:47 | |
*** javahorn has left #ingres | 04:47 | |
atrofast | Are there any Ingres benchmarks similar to that? | 05:07 |
pboro | nope... if I had some spare time, I would be glad to make one | 05:08 |
*** cthibert has joined #ingres | 05:10 | |
atrofast | just something that would be interesting at this point, not resting any key decisions on it or anything :) | 05:11 |
pboro | it would be interesting to see the effect of cache_dynamic = ON too | 05:11 |
atrofast | yes agreed | 05:15 |
*** rossand has joined #ingres | 05:30 | |
*** ChanServ sets mode: +o rossand | 05:30 | |
*** Dejan has joined #ingres | 06:03 | |
*** Mud has joined #ingres | 06:07 | |
Dejan | hello guys | 06:08 |
*** mull has quit IRC | 06:11 | |
*** Alex| has quit IRC | 06:18 | |
*** Alex| has joined #ingres | 06:38 | |
*** ChanServ sets mode: +o Alex| | 06:38 | |
*** mull has joined #ingres | 06:42 | |
*** toumi01 has joined #ingres | 06:48 | |
*** cthibert has left #ingres | 07:35 | |
*** cthibert has joined #ingres | 07:37 | |
*** Alex| has quit IRC | 07:57 | |
*** toumi01 has quit IRC | 08:16 | |
*** MikeT has joined #ingres | 08:23 | |
*** MikeT is now known as Guest37801 | 08:24 | |
*** Guest37801 has left #ingres | 08:24 | |
*** toumi01 has joined #ingres | 08:25 | |
*** Dejan has quit IRC | 09:28 | |
*** Alex| has joined #ingres | 09:36 | |
*** ChanServ sets mode: +o Alex| | 09:36 | |
*** grantc has quit IRC | 10:49 | |
*** KermitTheFragger has quit IRC | 10:54 | |
*** thiagomz has joined #ingres | 10:57 | |
thiagomz | Guys... I am trying to migrate a 2.6 legacy database to 10.0 Community edition.... I got a lot of this : E_CO003F COPY: Warning: 51543 rows not copied because duplicate key | 10:58 |
thiagomz | detected | 10:58 |
thiagomz | Anyone has i ideia why ? | 10:58 |
*** rossand has quit IRC | 11:00 | |
*** rossand has joined #ingres | 11:02 | |
*** ChanServ sets mode: +o rossand | 11:02 | |
*** Alex| has quit IRC | 11:41 | |
*** toumi01 has left #ingres | 11:57 | |
*** toumi01 has joined #ingres | 11:58 | |
*** toumi01 has quit IRC | 12:06 | |
*** toumi01 has joined #ingres | 12:06 | |
*** toumi01 has quit IRC | 12:10 | |
*** toumi01 has joined #ingres | 12:10 | |
*** thiagomz has left #ingres | 13:46 | |
*** cthibert has left #ingres | 14:15 | |
*** mull has quit IRC | 14:27 | |
*** atrofast has quit IRC | 14:55 | |
*** Mud has quit IRC | 14:58 | |
*** toumi01 has quit IRC | 15:55 | |
*** mull has joined #ingres | 16:31 | |
*** cytrinox_ has joined #ingres | 19:21 | |
*** cytrinox has quit IRC | 19:21 | |
*** cytrinox_ is now known as cytrinox | 19:21 | |
*** rossand has quit IRC | 19:40 | |
*** mull has quit IRC | 19:52 | |
*** Alex| has joined #ingres | 22:37 | |
*** ChanServ sets mode: +o Alex| | 22:37 | |
*** Mud has joined #ingres | 22:42 | |
*** Alex| has quit IRC | 23:26 | |
*** Alex| has joined #ingres | 23:39 | |
*** ChanServ sets mode: +o Alex| | 23:39 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!