| 00:56.13 | *** join/#storm keppla1 (n=keppla@i577B1FAE.versanet.de) |
| 00:56.18 | keppla1 | hi all |
| 01:08.12 | jkakar | Hello keppla1. |
| 01:08.21 | keppla1 | hi |
| 01:10.19 | keppla1 | i like to do some validation of my objects before commiting them to the database |
| 01:10.43 | keppla1 | kinda rails-style: is the customer name given, is the password sound |
| 01:11.11 | keppla1 | my current approach is to hook some homebrewn in __storm_pre_flush__ |
| 01:11.34 | keppla1 | is that a acceptable way, or am i missing something? |
| 01:12.24 | jkakar | keppla1: You want to use a validator. One sec, lemme find an example. |
| 01:12.38 | keppla1 | thanks |
| 01:14.35 | jkakar | keppla1: Okay, I remember how it works now. :) |
| 01:14.41 | keppla1 | :) |
| 01:14.57 | jkakar | keppla1: You define the validator key for the column you want to apply programmatic constraints to. |
| 01:15.27 | jkakar | keppla1: For example, title = Unicode(validator=my_validator_function) |
| 01:15.50 | keppla1 | oh, cool. that looks kind of obvious, i am sorry for not finding it myself |
| 01:16.04 | keppla1 | thank you! |
| 01:16.05 | jkakar | keppla1: my_validator_function has a signature like my_validator_function(obj, attr, value). It needs to return the value to use or raise an exception if the value is invalid. |
| 01:16.41 | jkakar | keppla1: You can see the code/docstring for it in storm/variables.py:128 (part of Variable.__init__). |
| 01:16.54 | keppla1 | thanks |
| 01:17.26 | jkakar | My pleasure. |
| 01:20.18 | keppla1 | how cool is that, i even can do implicit conversions, like "1.4" to 1.4 (maybe not so pythonic, but in my case kinda pragmatic) |
| 01:20.48 | jkakar | keppla1: Yeah, it was specifically designed for that kind of thing. :) |
| 01:20.57 | keppla1 | cool |
| 01:21.08 | keppla1 | ah, looking again i remember why i didnt find it: i just searched for help via python console: help Unicode, and the validator parameter is not mentioned there. Is there a better place to look, besides the tutorail and the manual |
| 01:21.22 | keppla1 | (ok, and "use the source", which i guess is what you did) |
| 01:21.28 | jkakar | keppla1: Sadly we're a bit documentation specific. |
| 01:21.45 | jkakar | keppla1: So yeah, the source is a good place to look. We do try to be careful and provide decent docstrings. |
| 01:22.02 | jkakar | keppla1: Also, feel free to ask questions here or on the mailing list. Here will usually get you a more timely response than the mailing list. |
| 01:22.06 | keppla1 | excusable for a framework like this, i guess |
| 01:22.13 | keppla1 | thanks again! |
| 01:22.37 | jkakar | keppla1: It's the usual issue of caring and wanting to do a good job, but not having quite enough time and having to make decisions that focus on pragmatism rather than completeness. |
| 01:23.11 | jkakar | keppla1: Er, I didn't mean "documentation specific", I meant "documentation deficient". :) |
| 01:23.22 | keppla1 | i absolutely like it (the comment was supposed to be a compliment, maybe i am too german to sound nice) |
| 01:23.34 | jkakar | keppla1: LOL! |
| 01:23.43 | jkakar | keppla1: FWIW, I always find German's very friendly. :) |
| 01:23.50 | keppla1 | thanks |
| 01:27.11 | *** part/#storm keppla1 (n=keppla@i577B1FAE.versanet.de) |
| 01:52.35 | *** join/#storm shaunm (n=shaunm@c-98-212-133-244.hsd1.il.comcast.net) |
| 04:55.45 | *** join/#storm jukart (n=jukart@d91-128-122-26.cust.tele2.at) |
| 04:59.56 | *** join/#storm keppla1 (n=keppla@i577B1FAE.versanet.de) |
| 05:00.01 | keppla1 | hi all |
| 05:06.47 | jamesh | hello |
| 05:10.11 | keppla1 | i am searching for some hints on "eager loading" |
| 05:11.17 | keppla1 | for example, i have tables a, b, c, and every a has many bs, and every b has many cs |
| 05:11.41 | keppla1 | i have to display a hierarchical list of a, b and cs |
| 05:11.59 | keppla1 | and i would like to get them all in one query |
| 05:14.54 | keppla1 | like "select a.*, b.*, c.* from a, b, c where a.id = b.a_id and b.id = c.b_id", id beeing the primary key |
| 05:15.23 | jamesh | so, you can do store.find((a, b, c), a.id == b.a_id, b.id == c.b_id) |
| 05:15.35 | jamesh | and the results will be tuples of those three tables |
| 05:16.26 | jamesh | that would issue almost exactly the same SQL as you suggest. |
| 05:20.25 | keppla1 | if i issue such a statement, and use the ReferenceSet of a to access bs, will that not hit the database again? |
| 05:21.10 | keppla1 | its about transparency, i would like my controllers to be ignorant of how to get the stuff from the model |
| 05:23.55 | keppla1 | hm, it looks that my experiments which led to this conclusion where flawed |
| 05:28.37 | keppla1 | ok, but with the corrected experiment, too, it looks like i cannot use the objects' referencesets without hitting the database |
| 05:31.21 | keppla1 | in rails terms, i am looking for something like "A.find(:all, :include => [ { :b => :c } ])" |
| 05:35.49 | jamesh | the reference set will issue queries, unfortunately. |
| 05:38.31 | jamesh | for some parts of Launchpad where it showed up on the profiles, we used a pattern of wrapping the model objects and overriding the ReferenceSet attributes with info from the query |
| 05:38.49 | jamesh | the lazr.delegates package makes that pretty easy |
| 05:40.57 | keppla1 | thank you, ill try that |
| 07:03.16 | *** part/#storm keppla1 (n=keppla@i577B1FAE.versanet.de) |
| 07:22.50 | *** join/#storm jukart (i=lovely@81.189.156.94) |
| 07:42.21 | *** join/#storm goschtl (n=goschtl@p5B0BF983.dip.t-dialin.net) |
| 08:39.39 | *** join/#storm seiflotfy (n=seif@41.234.24.162) |
| 09:12.18 | *** join/#storm seiflotfy (n=seif@41.234.24.162) |
| 12:31.25 | *** join/#storm vvinet (n=vince@smtpin.revolutionlinux.com) |
| 12:46.18 | *** join/#storm artista_frustrad (n=artista_@201-35-2-242.ctame704.dsl.brasiltelecom.net.br) |
| 13:22.01 | *** join/#storm adeuring (n=abel@p4FC5629E.dip.t-dialin.net) |
| 13:23.34 | adeuring | If I have a storm expression for a WHERE clause, is there an easy way to convert this expression into the SQL string? |
| 13:27.44 | therve | adeuring: probably something like "storm.expr.compile(expr)" |
| 13:31.14 | adeuring | thanks, I got that far :) Problem is that constants used in things like "table.column == 'value'" are not included in the result. You get instead "table.column = ?", and 'value' is stored in state.parameters, when you call state = storm.expr.State; compile(expression, state) |
| 13:32.00 | radix | yes, and that's because that's how storm sends the expression to the database |
| 13:32.08 | radix | it never interpolates them into the string of the expression |
| 13:32.14 | therve | adeuring: you don't want to interpolate strings |
| 13:32.47 | adeuring | therve: yeah, I know that is "bad habit", but I have to do this nevertheless is my case.... |
| 13:33.01 | therve | adeuring: why? |
| 13:33.43 | therve | it's not bad habit, it's just bad |
| 13:34.22 | adeuring | I have a set of existing Storm expressions and I want to use them in extisting code where the WHERE clause is generated as a plain string, without placeholders for parameters. |
| 13:35.20 | therve | adeuring: there is nothing in storm that will help you replace the parameters |
| 13:35.43 | radix | is it PHP, or something? |
| 13:36.02 | radix | that's the only database interface that I can think of off the top of my head that is stupid enough to not support bind parameters |
| 13:36.46 | radix | dbapi allows you to do conn.execute("select * from bar where a = ? and b = ?", (a, b)) |
| 13:38.00 | adeuring | radix: It's code that was originally based on SQLObject, and there you can have situations where it is hard to create complex WHERE expressions without sometimes using raw SQL strings |
| 13:48.42 | *** join/#storm seiflotfy2 (n=seif@41.234.24.162) |
| 14:53.27 | *** join/#storm artista_frustrad (n=artista_@201-35-2-242.ctame704.dsl.brasiltelecom.net.br) |
| 15:04.43 | *** join/#storm flaco__ (n=flaco@190.8.120.169) |
| 15:07.10 | flaco__ | hey all... I got a table with 2 datetime fields A and B... how can I find the row where A >= now() and now() <= B ?? |
| 15:16.43 | radix | store.find(Class, Class.A >= datetime.now(), datetime.now() <= Class.B) |
| 15:30.39 | jamesh | or use SQL('now()') |
| 15:30.52 | jamesh | if you want database transaction time rather than app time |
| 15:34.16 | *** join/#storm andrea-bs (n=andrea@ubuntu/member/beeseek.developer.andrea-bs) |
| 15:54.38 | *** join/#storm shaunm (n=shaunm@proxyserver.wolfram.com) |
| 15:58.15 | *** join/#storm artista_frustrad (n=artista_@201-35-2-242.ctame704.dsl.brasiltelecom.net.br) |
| 17:33.04 | *** join/#storm fcorrea (n=fcorrea@189.62.24.212) |
| 18:04.30 | *** join/#storm seiflotfy1 (n=seif@41.234.24.162) |
| 20:35.47 | fcorrea | hello there |
| 20:36.10 | jkakar | Hello fcorrea. |
| 20:36.47 | fcorrea | I read that in the past people used to have problems with storm and sqlite having "database locked" issues. I also read that this problem was solved..but it's happening |
| 20:36.52 | fcorrea | jkakar: hey :) |
| 20:37.18 | jkakar | fcorrea: Are you trying to access it from more than one thread or with more than one Store? |
| 20:38.16 | fcorrea | jkakar: It's a demo web app. And on every "view" there is a create_database |
| 20:38.37 | jkakar | fcorrea: Are you closing the store when you're done with it? |
| 20:39.03 | fcorrea | jkakar: that's what I was going to ask. Is there a closer? :) |
| 20:39.14 | jkakar | fcorrea: I don't use Storm with SQLite, so I'm mostly asking questions that seem obvious... I don't really understand the issues that lead to database locking issues. :( |
| 20:39.31 | jkakar | fcorrea: Yes, call store.close(). |
| 20:39.38 | jkakar | fcorrea: It closes the underlying database connection. |
| 20:39.57 | jkakar | fcorrea: Also, SQLite isn't thread-safe, so if you're accessing it from multiple threads you're likely going to run into trouble. |
| 20:40.43 | fcorrea | jkakar: right. I didn't have this problems with pgsql. So I think this is a corner case only related to sqlite |
| 20:41.32 | fcorrea | jkakar: I was asking if there was a closer because I never needed to close it using z2 and z3 |
| 20:44.35 | jkakar | fcorrea: Were you using ZStorm? If so, it would keep Store's alive for you across request boundaries (which you want, because establishing database connections for each request can be expensive). |
| 20:44.43 | *** join/#storm andym (n=andym@plone/andym) |
| 20:44.55 | jkakar | andym: Hey hey, welcome to our little party. :) |
| 20:45.03 | andym | hey jamu |
| 20:45.19 | andym | how was the holiday? |
| 20:45.59 | jkakar | andym: Absolutely fantastic, couldn't ask for more. Came home on Friday, feeling very refreshed. Happy to be back at work and getting back into the swing of things now. Four weeks is a long time to be away from home. :) |
| 20:46.06 | fcorrea | jkakar: yeah, I use IZStorm. But I also use middlestorm (which is a wsgi middleware that will always give me a safe store) |
| 20:46.18 | jkakar | fcorrea: Ah, cool. |
| 20:46.35 | andym | jkakar: great! |
| 20:46.45 | andym | you could ask for more. You could get more holiday :) |
| 20:46.49 | fcorrea | jkakar: this is mostly related to sqlite. No worries |
| 20:46.50 | jkakar | andym: How's the canoe? |
| 20:46.53 | jkakar | andym: Heh :) |
| 20:47.08 | andym | great, went out last night after work, saw seals and herons, the girls were thrilled |
| 20:47.13 | jkakar | Nice! |
| 20:47.51 | andym | got nothing to do with storm, just saw chat on #repoze :) |
| 20:48.34 | fcorrea | andym: aw..you're here too |
| 20:48.37 | fcorrea | ;) |
| 20:48.43 | andym | fcorrea: stalking you :) |
| 22:37.55 | *** join/#storm sidnei (n=sidnei@plone/dreamcatcher) |
| 23:57.49 | *** join/#storm andym (n=andym@plone/andym) |