| 00:43.20 | thumper | anyone awake? |
| 00:44.31 | therve | thumper: yes? |
| 00:44.36 | thumper | hey |
| 00:44.43 | thumper | I have a weird query I want |
| 00:45.02 | thumper | right now I'm loading a (BranchMergeProposal, Branch) tuple for each row |
| 00:45.15 | thumper | and trying to find the newest merge proposal for any given source branch |
| 00:45.19 | thumper | is that easy to do in storm? |
| 00:45.31 | therve | I guess? |
| 00:45.43 | thumper | I can't mental it out |
| 00:46.07 | therve | result = store.find((BranchMergeProposal, Branch), BranchMergeProposal.branch_id == Branch.id) |
| 00:46.20 | therve | result.order_by(BranchMergeProposal.creation_time) |
| 00:46.46 | thumper | therve: however there could be many merge proposals for a given branch |
| 00:46.51 | thumper | therve: and I just want the most recent |
| 00:47.00 | therve | ah ok, sorry |
| 00:47.03 | therve | I got your question wrong |
| 00:47.51 | therve | so I guess we need to find out how to do it in SQL first :) |
| 00:48.48 | therve | it looks like a GROUP BY/HAVING, but it's tricky |
| 00:50.14 | jamesh | if you don't mind selecting for the branches first, you could do something like: SELECT * FROM BranchMergeProposal WHERE (branch, creation_time) IN (SELECT branch, MAX(creation_time) from BranchMergeProposal WHERE branch IN (1, 2, 3, ...) GROUP BY branch) |
| 01:04.27 | fcorrea | hello there. Does anyone know why a resultset with a count of 1, wont work if .one() is called? Check this error: http://pastie.org/828215 |
| 01:08.01 | jamesh | fcorrea: what is "Content" in that paste? |
| 01:08.26 | fcorrea | jamesh: Content is a storm class |
| 01:08.45 | jamesh | how does it relate to the content.type class? (which appears to be Folder) |
| 01:09.29 | fcorrea | for any record in the Content table, you have a content.content_id reference |
| 01:10.25 | jamesh | it looks like you're effectively doing: SELECT Folder.* FROM Folder, Content WHERE Content.content_id = 2 |
| 01:10.38 | jamesh | which seems weird |
| 01:11.18 | fcorrea | this is the generated sql: [23:10:31.972651] EXECUTE: 'SELECT folder.content_id FROM content, folder WHERE content.content_id = %s', (2,) |
| 01:11.39 | fcorrea | and it fails at that query |
| 01:11.50 | jamesh | right. So you're going to retrieve every folder record. |
| 01:12.11 | fcorrea | mmmm |
| 01:13.08 | jamesh | if you want to retrieve one folder record, then include a constraint involving folder. |
| 01:13.22 | jamesh | if you want the content record, then ask for that class instead. |
| 01:13.33 | fcorrea | right |
| 01:13.36 | fcorrea | lemme try |
| 01:13.42 | fcorrea | I thought it was right |
| 01:18.22 | fcorrea | jamesh: you're definitely right. "Mistakes were made" according to bush |
| 01:18.23 | fcorrea | http://pastie.org/828227 |
| 01:21.12 | fcorrea | jamesh: ty much |
| 02:20.07 | *** join/#storm shaunm (~shaunm@c-98-212-133-244.hsd1.il.comcast.net) |
| 03:26.10 | *** join/#storm vvinet (~vince@modemcable244.133-202-24.mc.videotron.ca) |
| 04:23.08 | *** join/#storm jkakar (~jkakar@12.154.28.10) |
| 05:04.06 | *** join/#storm vvinet (~vince@modemcable244.133-202-24.mc.videotron.ca) |
| 06:16.26 | *** join/#storm jukart (lovely@81.189.156.94) |
| 07:24.16 | *** join/#storm stub (~stub@ppp-58-8-14-172.revip2.asianet.co.th) |
| 07:24.16 | *** join/#storm stub (~stub@canonical/launchpad/stub) |
| 07:24.42 | *** part/#storm stub (~stub@canonical/launchpad/stub) |
| 07:25.27 | *** join/#storm stub (~stub@ppp-58-8-14-172.revip2.asianet.co.th) |
| 07:25.28 | *** join/#storm stub (~stub@canonical/launchpad/stub) |
| 08:18.14 | *** join/#storm jukart_ (lovely@81.189.156.94) |
| 10:29.07 | *** join/#storm maxb_ (~maxb@jabberwock.vm.bytemark.co.uk) |
| 10:33.40 | *** join/#storm salgado-afk (~salgado@canonical/launchpad/salgado) |
| 10:56.58 | *** join/#storm jml (~jml@mumak.net) |
| 10:57.02 | *** join/#storm jml (~jml@canonical/launchpad/jml) |
| 12:08.13 | *** join/#storm keppla (~keppla@i577B2A73.versanet.de) |
| 13:17.29 | *** join/#storm vvinet (~vince@smtpin.revolutionlinux.com) |
| 13:51.06 | *** join/#storm jkakar (~jkakar@12.154.28.10) |
| 14:00.45 | *** join/#storm niemeyer (~niemeyer@12.154.25.188) |
| 16:42.00 | *** join/#storm danilos (~danilo@canonical/launchpad/danilos) |
| 17:57.50 | vds | anyone has time to help me with a problem with ReferenceSet? http://pastebin.com/d11017373 |
| 17:59.24 | therve | hum... |
| 17:59.46 | vds | therve: hello |
| 17:59.53 | therve | hi vds |
| 18:00.27 | vds | therve: I had the same problem when that reference was a Reference and not a ReferenceSet but solved defining on_remote |
| 18:00.44 | vds | equal True |
| 18:01.11 | vds | from the code looks like the second key is always considered remote |
| 18:01.22 | therve | yeah |
| 18:01.25 | vds | is it a cache problem or something else? |
| 18:03.04 | therve | vds: what version of storm are you using? |
| 18:07.51 | vds | therve: 0.16.0.99 |
| 18:08.32 | therve | vds: any idea which bzr revision it may be? |
| 18:09.28 | therve | vds: so if it's not > r342, I would try in case of, as some changes happened |
| 18:09.46 | *** join/#storm fcorrea (~fcorrea@201-75-152-196-nd.cpe.vivax.com.br) |
| 18:09.47 | therve | vds: in the mean time, using self.devices.remove(device) should do the tricky |
| 18:09.49 | therve | trick |
| 18:09.53 | vds | therve: unfortunately no |
| 18:10.14 | therve | vds: self.devices.remove doesn't work? |
| 18:10.42 | vds | therve: that one was about the version but the answer is still no |
| 18:10.48 | therve | heh |
| 18:10.54 | vds | :) |
| 18:11.01 | therve | you still need to call store.remove(device), right? |
| 18:11.06 | vds | because the foreign key is not supposed to be Null |
| 18:12.03 | therve | humph :/ |
| 18:12.26 | vds | but if I understand correctly the entry in the db is removes |
| 18:12.47 | vds | removed |
| 18:13.23 | vds | but as the User instance is not touched the cache doesn't invalidate it |
| 18:13.52 | vds | I've tried store.invalidate(user) but didn;t work and I'm not even sure that makes any sense |
| 18:21.14 | *** join/#storm jukart (~jukart@d86-32-163-62.cust.tele2.at) |
| 20:04.25 | jkakar | vds: Did you sort out the issue? Sorry to be MIA... had some calls, then lunch, etc. |
| 22:44.44 | vds | jkakar: nope still trying to figure it out |
| 22:45.22 | jkakar | hugs vds |
| 22:49.10 | vds | :) |
| 22:49.58 | vds | it's definitely strange I though that using ReferenceSet instead of Reference would have been easier, but looks like I'm wrong! :) |
| 23:01.22 | jkakar | vds: Well, ReferenceSet is for 1->M relationships, so I'm not sure it's the right fit...? |
| 23:02.47 | vds | jkakar: should be fine |
| 23:03.02 | vds | I mean that's what I'd like to have |
| 23:22.59 | jkakar | vds: Ah, okay. |
| 23:25.51 | vds | jkakar: is there a way to invalidate the cache? |
| 23:26.19 | jkakar | vds: Yep, store.invalidate() |
| 23:27.03 | vds | jkakar: no arguments? |
| 23:27.35 | jkakar | vds: Nope. |
| 23:27.50 | jkakar | vds: Do you want to invalidate a particular object or the entire cache? |
| 23:28.17 | jkakar | vds: If you want to invalidate an object you can do store.invalidate(obj). |
| 23:28.42 | vds | jkakar: uhm I've tried that but didn;t work |
| 23:31.51 | jkakar | vds: Hmm. |
| 23:32.19 | vds | let me try it again |