pdf processing challenge

[bfuzze_98]

I was wondering if anyone has any experience processing data or extracting data from pdf documents?

I have a group that uses an online system to place orders. Documentation of the order is output in pdf format. The same group uses another (separate) online utility to log and track purchases. I'd like to create a script that can process these pdf's for automated upload into the database the utility extracts from.

I found a feature in Acrobat 6.0 called batch processing. One of the predefined processes is to convert a selection of (multiple/single) files to rtf's (rich-text). That is half the battle right there. Unfortunately, I don't see a php function comparable to this Acrobat feature (not surprising since most php functions are for generating pdfs not breaking them down). I'm just beginning to look into this, but my initial/backup plan is to manually convert these saved pdf receipts into rtfs and then process the rtfs using reg exp.

Does anyone have any other ideas/methods of processing these documents? Is there a way (script) to automate this conversion from pdf to rtf? Any suggestions, comments, or potential pitfalls would be appreciated.

Thanks!

[etully]

There are probably 100 ways to do that. I've seen at least 20 myself.

But before you go down that road, you really need to examine the work flow here. To prevent errors (and to skip headaches like the one you're having now), the correct way to build this system is to have the original ordering system output the order twice.

It can still output a PDF for the users... but then at the same time, it should be writing the details of the order to a database or, at the very least, an XML or a CSV file.

I'm talking about good programming. Good system design. It's better for the client and it's better for you. Someday, someone is going to ask that the PDF's get formatted differently and you will have to modify your PDF data extraction script again - if it can even be done. (Imagine if they said, we want the order total displayed in graphics, not text. Now your import script can't work no matter what).

The system you are designing is called a kludge - and I understand the appeal, it's the fastest way to be done. But trust me, after 30 years of programming, the simplest solution is always best. Tell the client that they can pay $300 more today to do it right... or $8000 more in five years from now to clean up the mess.

[bfuzze_98]

etully,

Thanks for your input, however, I may have neglected to mention that I do not have control over the initial data output format. The purchasing is done through a 3rd party application (beyond my control) and I am working with just one small group within a very large organization. I agree the ideal situation would be a directly write to the database in addition to the pdf output.

In addition, while I recognize how quickly technology and demands seem to change in this industry, this organization just upgraded from a DOS based accounting system:eek: to SAP:quiet: so they're not exactly quick to change (not to mention the transition for such a large org is an immense undertaking).

That said, I just want to accomplish the task at hand, quick/dirty/how ever I can... and move on. It may not be pc programming etiquete, but it's my only option at the moment.

You said you know of some good ways to process these pdfs? Can you point me in the right direction?

Thanks!