Comments on: Dynamic SQL querys in ActionScript http://www.angrysprite.com/2008/04/04/6/ trace(actionscriptRants[Math.round(Math.random() * (actionscriptRants.length - 1))]); Tue, 06 Jan 2009 00:20:13 +0000 http://wordpress.org/?v=2.5 By: Matthew http://www.angrysprite.com/2008/04/04/6/#comment-3 Matthew Tue, 22 Apr 2008 21:21:38 +0000 http://www.angrysprite.com/?p=6#comment-3 Thank you for the feedback. I agree that passing the objects/arrays in is not very elegant. I went with that for the now because this allow for me to move forward with my testing and get the big picture working. This is very much a work in progress. Thank you for the feedback.
I agree that passing the objects/arrays in is not very elegant. I went with that for the now because this allow for me to move forward with my testing and get the big picture working.

This is very much a work in progress.

]]> By: Gareth http://www.angrysprite.com/2008/04/04/6/#comment-2 Gareth Fri, 18 Apr 2008 16:07:07 +0000 http://www.angrysprite.com/?p=6#comment-2 I really like what you've done so far. It would certainly make the SQL creation much easier. The Adobe application, Employee Directory, (I think) uses the Cairngorm model and moves all of the SQL calls into separate AS classes. These are then passed to a SQLmanager which invokes the necessary SQL call, and passes the results back to the calling page. This somewhat separates the SQL code from the AS code, but requires one page for each SQL statement, which can get a little overwhelming if you have many calls to the SQLite dB. The only thing that I did not like *too* much is passing in the objects into the filter array. It's all a little mysterious as to what to pass in to the array. A new user would not know exactly what to pass in without either the code or an API. Here's something that I'm not sure would work completely, but might or might be a step in the right direction :) How about creating a FilterableObject? This object would have public methods for each of the properties within your case statements in your parseTerms method. Each of the methods would return "this" (the FilterableObject), so you could do new FilterableObject().isNotNull( 'insider_id' ).iExact( 'name', 'adsf' ) etc., and chain everything together. Eclipse would then be able to generate the "hints" as you're typing and the array that is passed in wouldn't be as mysterious to a new user. Anyway, this will be a really useful piece of code once it's finished, and I look forward to seeing it completed. I really like what you’ve done so far. It would certainly make the SQL creation much easier. The Adobe application, Employee Directory, (I think) uses the Cairngorm model and moves all of the SQL calls into separate AS classes. These are then passed to a SQLmanager which invokes the necessary SQL call, and passes the results back to the calling page. This somewhat separates the SQL code from the AS code, but requires one page for each SQL statement, which can get a little overwhelming if you have many calls to the SQLite dB.

The only thing that I did not like *too* much is passing in the objects into the filter array. It’s all a little mysterious as to what to pass in to the array. A new user would not know exactly what to pass in without either the code or an API.

Here’s something that I’m not sure would work completely, but might or might be a step in the right direction :)

How about creating a FilterableObject? This object would have public methods for each of the properties within your case statements in your parseTerms method. Each of the methods would return “this” (the FilterableObject), so you could do
new FilterableObject().isNotNull( ‘insider_id’ ).iExact( ‘name’, ‘adsf’ ) etc., and chain everything together. Eclipse would then be able to generate the “hints” as you’re typing and the array that is passed in wouldn’t be as mysterious to a new user.

Anyway, this will be a really useful piece of code once it’s finished, and I look forward to seeing it completed.

]]>