Lakers were trounced -- Boo!
Frustrated that I couldn't manipulate a large XML file in CFMX, I
decided to write a Java program to read a text file, line by line.
Didn't need to --- found one by Daryl Banttari, at:
http://www.macromedia.com/devnet/server_archive/articles/
leveraging_java_classes_cf.html
This works fine -- takes about 3 seconds to read the 3.8 meg, 85,000
line file.
Interesting!
1) If it is so easy, why not implement it CFML? -- the CFMX/Java
hybrid solution means extra work when you install or maintain the app
-- you have to deal with Java source and class files.
2) The routine points out another deficiency in CFML (or at least
CFML's interface to other languages, like Java) -- CFML cannot
represent/recognize a Null! The programmer had to provide a
work-around.
Don't get me wrong, I am happy to find a solution, and being able to
interface Java is great!
But, IMO, this should not be necessary (to read a large file,
line-by-line) for a language as rich and powerful as CFML!
Dick
On Jun 15, 2004, at 9:08 PM, Dick Applebaum wrote:
>
> OK!
>
> 1) I can't readLn a file in CF so I read the entire 3.8 meg xml file
> into RAM
>
> 2) XmlTransform runs out of memory. so I can't use xls to manipulate
> the file
>
> 3) XmlSearch runs out of memory so I can't use Xpath to manipulate the
> file
>
> 4) I can CFexecute a Perl program to parse/render the file in 3-4
> seconds
>
> 5) I can convert the 3,8 Meg, 85,000 line file to an in-memory array
> (Now I have the orig file and the array representation of this file in
> memory at the same time..
>
> 6) Now, I can line-by-line (array-element-by-array-element) process
> this file.
>
> What's wrong with this picture --- AFAICT -- everything!
>
> CF is the wrong tool!
>
> Thought I'd never say that!
[Todays Threads] [This Message] [Subscription] [Fast Unsubscribe] [User Settings] [Donations and Support]