| 01:04.05 | *** join/#storm paranoidi (n=paranoid@av41.adsl.tnnet.fi) [NETSPLIT VICTIM] |
| 01:04.05 | *** join/#storm munga (n=munga@81.194.35.9) [NETSPLIT VICTIM] |
| 02:38.31 | *** join/#storm bigdog1 (n=scmikes@72-197-8-8-arpa.cust.cinci.current.net) |
| 03:55.44 | *** join/#storm MFen (n=cdodt@demo.goonmill.org) |
| 03:56.48 | MFen | what's the recommended way to execute an sql script for setting up? |
| 03:57.11 | MFen | i have a bootstrap script that runs my create table from an sql script. this seems like a straightforward, sane way to do things |
| 03:57.37 | MFen | but then i need to do the same thing to an in-memory database. so i have to run the sql from python. and .execute() doesn't accept multiple statements |
| 03:58.06 | MFen | this is for writing unit tests |
| 03:58.26 | MFen | i don't see how to do that except to get rid of my nice .sql file and replace it with a list of statements in python code |
| 04:37.30 | *** part/#storm MFen (n=cdodt@demo.goonmill.org) |
| 04:37.37 | *** join/#storm MFen (n=cdodt@demo.goonmill.org) |
| 04:53.01 | *** join/#storm sidnei (n=sidnei@plone/dreamcatcher) |
| 05:12.40 | *** join/#storm jukart (n=jukart@d91-128-122-124.cust.tele2.at) |
| 05:16.13 | *** join/#storm noodles775 (n=miken@e178214017.adsl.alicedsl.de) |
| 05:19.50 | *** join/#storm sidnei_ (n=sidnei@201-40-159-6.cslce701.dsl.brasiltelecom.net.br) |
| 05:23.26 | *** join/#storm jukart_ (n=jukart@d91-128-122-124.cust.tele2.at) |
| 05:48.05 | *** join/#storm sidnei_ (n=sidnei@189-10-169-175.cslce701.dsl.brasiltelecom.net.br) |
| 06:14.35 | *** join/#storm jukart (i=lovely@81.189.156.94) |
| 07:01.59 | *** join/#storm sidnei__ (n=sidnei@200.203.78.39) |
| 07:24.26 | *** join/#storm thumper (n=tim@canonical/launchpad/thumper) |
| 07:49.23 | *** join/#storm goschtl (n=goschtl@p5B0BEF66.dip.t-dialin.net) |
| 07:51.43 | *** join/#storm sidnei__ (n=sidnei@201-66-186-231.cslce701.dsl.brasiltelecom.net.br) |
| 08:43.07 | *** join/#storm sidnei__ (n=sidnei@201-66-192-186.cslce701.dsl.brasiltelecom.net.br) |
| 09:19.06 | *** join/#storm sidnei__ (n=sidnei@201-66-131-157.cslce701.dsl.brasiltelecom.net.br) |
| 09:22.57 | *** join/#storm munin (n=munin@209.9.227.50) |
| 09:41.30 | *** join/#storm thumper (n=tim@canonical/launchpad/thumper) |
| 09:44.13 | *** join/#storm sidnei_ (n=sidnei@201.3.148.190) |
| 10:30.25 | *** join/#storm sidnei__ (n=sidnei@201.66.186.178) |
| 11:11.43 | *** join/#storm gord (n=gord@5ac3280b.bb.sky.com) |
| 11:39.52 | *** join/#storm niemeyer (n=niemeyer@200-138-37-135.ctame705.dsl.brasiltelecom.net.br) |
| 11:51.41 | *** join/#storm pepit678 (n=pepit678@161.67.213.5) |
| 11:52.30 | *** part/#storm pepit678 (n=pepit678@161.67.213.5) |
| 12:00.29 | *** join/#storm sidnei__ (n=sidnei@201-66-186-212.cslce701.dsl.brasiltelecom.net.br) |
| 12:04.34 | *** part/#storm sidnei__ (n=sidnei@201-66-186-212.cslce701.dsl.brasiltelecom.net.br) |
| 12:18.29 | *** join/#storm noodles775 (n=miken@e178199163.adsl.alicedsl.de) |
| 13:37.21 | jkakar | MFen: FWIW, we use a list of Python statements in Landscape to manage our schemas. |
| 13:57.46 | MFen | jkakar: i'm going to do something like "sqlite3 -init <(python sqlschema.py) my.db" so i can use the same script in both places |
| 13:58.08 | jkakar | MFen: Makes sense. |
| 13:58.33 | jkakar | MFen: For performance reasons, we create fresh schemas only when necessary; the rest of the time we run a bunch of DELETE statements to clear out the tables, between test runs. |
| 13:59.03 | jkakar | MFen: This requires that our application can function with completely empty databases (which it can). |
| 14:00.50 | *** join/#storm jukart (n=jukart@d91-128-122-124.cust.tele2.at) |
| 14:03.19 | MFen | jkakar: well, for tests it's easy enough to just create a blank in-memory database and run all the sql on it in every setup |
| 14:03.48 | MFen | but either way, i have to be able to do that at least once, which means having a list of sql statements instead of a nice sql script |
| 14:04.51 | jkakar | MFen: Yeah. "First make it work, then make it fast" is a good way to do things. :) |
| 14:06.58 | MFen | creating an in-memory database is very fast, actually |
| 14:08.33 | *** join/#storm sidnei (n=sidnei@plone/dreamcatcher) |
| 14:11.25 | jkakar | MFen: Yes, SQLite is nice for that. In our case, we use PostgreSQL. Creating schemas is not fast. :) |
| 14:11.42 | MFen | ah. |
| 14:11.52 | jkakar | You probably won't run into performance problems until you have hundreds of tests. |
| 15:09.50 | *** join/#storm jukart (n=jukart@d91-128-122-124.cust.tele2.at) |
| 15:34.21 | *** part/#storm sidnei (n=sidnei@plone/dreamcatcher) |
| 15:34.36 | *** join/#storm shaunm (n=shaunm@proxyserver.wolfram.com) |
| 15:37.01 | *** join/#storm jkakar_ (n=jkakar@S0106001ee57a879c.vc.shawcable.net) |
| 15:48.46 | MFen | jkakar_: at that point, wouldn't deletes also give me performance problems? |
| 15:49.29 | jkakar_ | MFen: In PostgreSQL, at least, DELETEing from all tables is faster than DROPing and CREATEing them from scratch. It's also faster than using TRUNCATE. |
| 15:50.00 | MFen | right, but in a sqlite memory database i doubt there's much difference |
| 15:50.17 | jkakar_ | Yeah, I could imagine that being the case. |
| 16:16.47 | *** join/#storm andrea-bs (n=andrea-b@ubuntu/member/beeseek.developer.andrea-bs) |
| 17:22.04 | *** join/#storm jkakar (n=jkakar@S0106001ee57a879c.vc.shawcable.net) |
| 18:10.35 | *** join/#storm andrea-bs (n=andrea-b@host56-27-dynamic.30-79-r.retail.telecomitalia.it) |
| 19:58.35 | *** join/#storm sidnei (n=sidnei@plone/dreamcatcher) |
| 20:31.49 | *** join/#storm mup (n=mup@li37-90.members.linode.com) |