| 00:00.19 | *** join/#storm wallflower (n=wallflow@ip-205-246-113-216.pool.grokthis.net) |
| 00:01.50 | *** join/#storm darkhorse` (n=admins@hill-d-029.resnet.purdue.edu) |
| 00:12.36 | *** join/#storm shaunm (n=shaunm@c-98-212-133-244.hsd1.il.comcast.net) |
| 00:23.13 | spats | hey all. I have a subselect formed using Select that does a group-by on an Alias of one of my tables. How do I now refer to the columns of its results in a find query, such that it will be included in the SQL as an implicit inner join? |
| 00:26.42 | spats | (correction: actually a ClassAlias) |
| 01:25.21 | spats | nm, found that expr.Column can be used to do this. |
| 02:05.41 | *** join/#storm vvinet (n=vince@modemcable101.133-202-24.mc.videotron.ca) |
| 02:21.06 | *** join/#storm therve (n=therve@xvm-14-10.ghst.net) |
| 05:54.29 | *** join/#storm radix` (n=r@wordeology.com) |
| 06:17.32 | *** join/#storm jukart (i=lovely@81.189.156.94) |
| 07:33.23 | *** join/#storm goschtl (n=goschtl@p5B0BFD92.dip.t-dialin.net) |
| 10:00.38 | *** join/#storm niemeyer (n=niemeyer@nat/canonical/x-1885e3c1cc2f5ffc) |
| 10:05.23 | cody-somerville | I have a table called BuildResults that has a column that is a DateTime field. I'm wondering what the easiest way it would be to calculate the number of rows per day. |
| 10:06.46 | therve | cody-somerville: do you need the number of rows for all days, or for a particular day? |
| 10:06.57 | cody-somerville | for all days |
| 10:07.01 | cody-somerville | I want to graph the data |
| 10:07.24 | therve | therve: a group by should do it then |
| 10:08.20 | therve | wow I talk to myself now |
| 10:08.48 | cody-somerville | How would it group the DateTime field? The field has a date and time, but I just care about the date |
| 10:08.49 | therve | cody-somerville: you should use a count(*) with a group by, casting your datetime as a date |
| 10:08.54 | cody-somerville | ah |
| 10:09.04 | cody-somerville | Is there an example on the wiki? |
| 10:09.38 | therve | that would surprise me |
| 10:13.04 | cody-somerville | So, what do I pass group_by? |
| 10:13.27 | cody-somerville | ah |
| 10:13.29 | cody-somerville | an expression |
| 10:13.43 | cody-somerville | how do I get storm to cast it? |
| 10:14.17 | therve | cody-somerville: are you using postgres? |
| 10:15.06 | cody-somerville | yup |
| 10:16.11 | therve | cody-somerville: the easiest way is probably to use date_trunc then |
| 10:16.27 | therve | you're defining a subclass of Function |
| 10:17.11 | therve | NamedFunc |
| 10:17.35 | therve | class DateTrunc(NamedFunc): name="date_trunc" |
| 10:17.51 | therve | then group_by(DateTrunc("day", mytimefield)) |
| 10:19.15 | cody-somerville | So, where would be the right place to put this logic? |
| 10:22.42 | therve | cody-somerville: what do you mean? |
| 10:22.46 | therve | in your code? :) |
| 10:23.03 | cody-somerville | Like, should I have it in model or should I have it in my controller? |
| 10:23.53 | therve | I'm sure I don't want to talk about that :) |
| 11:12.24 | *** join/#storm gord (n=gord@5ac7eaa6.bb.sky.com) |
| 12:07.41 | *** join/#storm deryck (n=deryck@samba/team/deryck) |
| 13:07.20 | cody-somerville | therve, do you think it would make sense to store the duration of the builds in the database so that I can just use a simple aggregation function? |
| 13:27.26 | therve | cody-somerville: to answer that question, you need to take into account the amount of data, the number of fields, the cost of the query, etc |
| 15:58.01 | *** join/#storm andrea-bs (n=andrea-b@ubuntu/member/beeseek.developer.andrea-bs) |
| 16:29.44 | *** join/#storm shaunm (n=shaunm@proxyserver.wolfram.com) |
| 16:45.22 | shaunm | anybody have any tips for managing storm's object cache? |
| 16:45.34 | shaunm | pulse is running out of memory when crawling large modules |
| 16:47.38 | therve | shaunm: that would be a good candidate for the GenerationalCache |
| 16:47.41 | sidnei | shaunm: we've landed a generational cache recently, which should help with that |
| 16:47.45 | sidnei | therve: beat me to it! |
| 16:50.01 | shaunm | recently? not in 0.14? |
| 16:50.12 | mzz | yes, yes |
| 16:50.25 | mzz | (iirc) |
| 16:50.36 | shaunm | anywhere I could read up on it? docs, ml archives, whatever |
| 16:51.38 | mzz | I came across it in http://bazaar.launchpad.net/~storm/storm/trunk/annotate/head%3A/NEWS but there's almost certainly better ways to get more info |
| 16:52.04 | therve | yeah, looking at the source code |
| 16:54.03 | shaunm | heh, all right |
| 16:57.48 | radix` | you just need to construct your store with it |
| 16:59.21 | shaunm | right, ok |
| 16:59.47 | shaunm | contemplates grabbing the latest storm from bzr |
| 17:00.06 | shaunm | sadly, I don't have bzr on this machine :( |
| 17:00.21 | mzz | this is not hard to fix! :P |
| 17:01.38 | shaunm | well, harder than it could be. but not terribly hard, I suppose |
| 17:02.03 | shaunm | I'm on an old machine. it's not in yum |
| 17:03.21 | mzz | just grab the .tar.gz and either install it into your homedir or run straight from the .tar.gz (which is a bit slower because you're missing the compiled extension modules but should still be perfectly acceptable for cloning a branch over http once) |
| 17:04.10 | mzz | assuming you do have a recent enough python |
| 17:04.50 | mzz | err, duh, you're running storm |
| 17:05.05 | shaunm | heh |
| 17:05.41 | therve | shaunm: you can also just get the class in the source browser and copy it in your project |
| 17:05.42 | shaunm | it's only python 2.4.4 |
| 17:06.09 | shaunm | therve: ooh, that's not a bad idea |
| 17:20.42 | shaunm | hmm, hanging on the very first select |
| 17:22.33 | shaunm | nevermind, had a stale lock |
| 17:30.05 | *** join/#storm sidnei (n=sidnei@plone/dreamcatcher) |
| 17:36.26 | shaunm | hooray! |
| 17:36.47 | shaunm | I didn't run out of memory processing gtk+ |
| 18:08.54 | radix` | shaunm: so all you had to do was switch to the generational cache? |
| 18:09.36 | shaunm | yes |
| 19:56.56 | *** join/#storm sidnei_ (n=sidnei@200-102-242-25.cslce701.dsl.brasiltelecom.net.br) |