10:07 <mynyml> is there a way to load a sql schema into an in memory sqlite db?

10:08 <Death_Syn> i'd do it via migrations

10:18 <mynyml> Death_Syn: that'd be ideal but i can't.

10:19 <mynyml> closest i can get is a sql schema dump

10:24 <bougyman> sqlite3 the.db < the.sql.dump ?

10:24 <bougyman> i'd assume that would work, though i've never tried.

10:35 <Death_Syn> he's trying to load it into a memory-based db

10:36 <bougyman> oh

10:36 <bougyman> wild

10:36 <exploid> Sequel["sql goes here"] ?

10:56 <mynyml> exploid: hm! looks like i was looking too hard. trying it now

11:10 <mynyml> exploid: works like a charm. thx

11:11 <mynyml> my specs are now 10x faster (letterally)

11:12 <pusewicz> what's the easiest way to process data in batches?

11:33 <jason^> is there a way to do an insert ignore with sequel?

12:28 <jeremyevans> jason^: http://sequel.rubyforge.org/rdoc-adapters/classes/Sequel/MySQL/DatasetMethods.html#M000190

13:35 <jason^> jeremyevans: thanks, i didn't realize that was in there

14:23 <bman> jeremyevans:hi

15:10 <tibbon> If I've set a column an attribute as unique => true, and then I try to insert something that's a duplicate, how do I ignore the error and keep going? The user doesn't need to know that I've already got the entry. Getting the "SQLite3::SQLException column FOO is not unique" error. I'm using Sinatra as my framework.

15:17 <exploid> tibbon: wrap your call in a begin rescue block and just ignore the exception?

15:18 <tibbon> exploid: Around the line that I'm actually trying to call the insert, not somewhere in the model right?

15:19 <tibbon> exploid: Got it, works perfectly. Thanks!

15:20 <exploid> tibbon: no problem

15:20 <exploid> tibbon: you might want to catch more than one exception type though, in case it would be an exception other than the one for duplicate record

15:20 <exploid> tibbon: not sure what the exceptions are

15:28 <jeremyevans> tibbon: You want to catch Sequel::DatabaseError, since that is what should be raised for all database exceptions

15:28 <jeremyevans> bman: Hi

15:28 <tibbon> jeremyevans: Stupid question, but what's the syntax of that line?

15:29 <exploid> raise Sequel::DatabaseError => exception_var

15:29 <jeremyevans> tibbon: begin; dataset.insert(...); rescue Sequel::DatasetError; nil; end

15:29 <tibbon> jeremyevans: perfect. Thank you much

15:30 <bman> jeremyevans: did something change for column maps... i get this error: undefined method `map' for nil:NilClass

15:37 <jeremyevans> bman: Can you pastie?

15:44 <bman> http://p.ramaze.net/23620

15:45 <bman> although the error is coming from a ramaze helper, it points to the columns.map... which is why i'm asking has something broken in sequel that is causing this... worked fine in previous version of sequel

15:45 <jeremyevans> bman: That should have only worked if you were using set_schema in the model

15:46 <bman> ah, yes... and i removed that, since you mentioned that it's discouraged to use set_schema

15:46 <jeremyevans> bman: I'm not sure what the purpose of that code is, but there's definitely a better way to do it.

15:46 <bman> ok

15:46 <jeremyevans> bman: What are you trying to do?

15:47 <bman> it's generating a html form from a model

15:47 <jeremyevans> OK. You can get the columns from Model.columns

15:47 <jeremyevans> bman: If you want column related information, use Model.db_schema

15:47 <bman> and it returns the column data type

15:48 <jeremyevans> bman: Model.columns.map{|c| Model.db_schema[c]} may get you close to what you want

15:48 <bman> i.e. bit, i make a checkbox, varchar make a input box, etc

15:48 <jeremyevans> bman: The :type entry in the resulting hash will be something like :string, :integer, etc.

15:49 <jeremyevans> bman: And using checkboxes for booleans is a bad idea unless the boolean is NOT NULL

15:49 <bman> what would you suggest

15:51 <jeremyevans> bman: For booleans that can be NULL, a select box with 3 options (NULL/empty, true, and false) is how I handle that

15:51 <jeremyevans> bman: The :allow_null entry in the hash shows whether nulls are allowed

15:52 <bman> hmmm, and your web customers/users base adapt to that pretty well?

15:53 <jeremyevans> bman: Generally these are admin forms, but I haven't had complaints from the users who have to use them.

15:54 <bman> jeremyevans: is there a hash that contains all the possible :types... i want to create a case statement for my forms generator

15:54 <bman> or do i have to look at the code?

15:54 <jeremyevans> bman: I don't believe there is currently.

15:55 <bman> are types based on the adapter, or are they standard across all adapters

15:55 <jeremyevans> bman: Database#schema_column_type is where you should look. I believe it is standard across adapters

15:55 <bman> k

15:56 <bman> you don't have a form generator up your sleeve somewhere do you :) ... (no need to re-invent)

15:57 <jeremyevans> bman: Scaffolding extensions has one you could probably use

15:57 <bman> so, i'd have to hack it out of your code since it's not ramaze 2009.04 ready?, correct?

15:57 <jeremyevans> bman: You might have to use quite a bit of that code though, since it's pretty abstracted in order to support both Sequel and AR

15:58 <jeremyevans> bman: The form generation shouldn't be specific to any web framework

15:58 <jeremyevans> bman: The issues with Ramaze 2009.04 are probably in the rendering. The form generation is the same on all 5 supported frameworks

15:59 <bman> k

16:09 <ddollar> wondering if someone can give me a hand.. i have an array like ['foo', 'bar', 'fred'] and i want to get a where clause of WHERE (col1 LIKE '%foo%' OR col1 LIKE '%bar%' OR col1 LIKE '%fred%')

16:13 <jeremyevans> ddollar: dataset.grep(:col1, %w'foo bar fred')

16:13 <ddollar> jeremyevans: awesome, that's way simpler than what i was trying to do :P

16:13 <ddollar> thanks

16:14 <jeremyevans> ddollar: Oops, that's not exactly what you want, though

16:14 <ddollar> do i need to add the %

16:14 <ddollar> 's?

16:14 <jeremyevans> ddollar: dataset.grep(:col1, %w'foo bar fred'.map{|x| "%#{x}%"})

16:15 <jeremyevans> ddollar: And you may want to escape the % inside x as well:

16:15 <jeremyevans> ddollar: dataset.grep(:col1, %w'foo bar fred'.map{|x| "%#{x.gsub('%', '%%')}%"})

16:16 <jeremyevans> ddollar: I'm assuming your database is fairly small?

16:16 <ddollar> yeah

16:16 <ddollar> very

16:16 <ddollar> only a few hundred rows

16:16 <jeremyevans> ddollar: Then performance shouldn't be an issue

16:17 <ddollar> yup.. it works great as manual sql, im just trying to convert the project to sequel so i can stop building sql by hand :)

16:26 <mynyml> anyone knows where i could find simple docs for mime type handling?

16:27 <mynyml> i'm seeing :provides in the source for HTTP_ACCEPT ...

16:27 <mynyml> just wondering if there's a simple way to handle mime types. basically i need to allow for xml and json

16:28 <jeremyevans> mynyml: This is #sequel :)

16:28 <mynyml> *sign*

16:28 <mynyml> *sigh*

16:28 <mynyml> sorry about that

19:30 <hgimenez> jeremyevans, hi

19:33 <hgimenez> jeremyevans, wondering if you had any thoughts on those patches

19:49 <kupa> jeremyevans: Just dumped 1.0.3 merb_sequel which adds rspec examples wrapped in transactions so model specs are fast(er)! :-), thanks Jeremy.

20:14 <jeremyevans> hgimenez: I'm sorry, I haven't had a chance to look at them yet

20:15 <jeremyevans> kupa: Thanks!

20:39 <jeremyevans> hgimenez: I merged your changes, thanks!

20:47 <hgimenez> jeremyevans, sorry, was gone. Thank you, that's great!