Import file with images

[JimmyLewis]

Hi,

I have many banking transactions and various scanned receipts (not every transaction has a receipt). I am trying to find a way to use the feature of "import file" and import all scanned receipts. The problem is that I want those receipts to correspond to the appropriate transaction. Is there any way to do this in a way that makes sense?

Thanks,

jimmy

[Fenton]

Well, that's hard to say. Probably not when you first Import them (unless you already have the exact file name or path in a FileMaker field). You can Import the file name and the file path. The question is whether there is information in the name or path which you could parse out, during a script, to tell you what they belong to.

[JimmyLewis]

Even naming the receipt manually would be less work. The problem with that is: what would I name the file? there are no fields that are unique enough for me to match them without actually finding the transaction (which is what i was trying to avoid). the dates are not always the same, the store name is always preceded by random numbers and letters, and the amounts alone are not enough to be unique.

So since it seems there is no other way, all i can think of is

to name the file based on the transaction id.

The problem is when I try to import, it shows the file name with its extension, even when the extension is hidden. how can i parse information from an import in such a case?

The only way i can think of is by doing it after the fact, using a calculation field, i.e. sending the filename to text field, then having the serial field with the relationship to the transaction be based on a calculation like left() and the number of characters in the serial number, which is constant.

is this how you would do it in such a scenario?

[Fenton]

It's kind of hard to say how I'd do it, as I don't know how the receipt scans are generated, nor what your folder structure is. But, as you can see, just having a bunch of files with no identifying structure or information is a problem. There's a number of possible approaches.

If you could name the file with the transaction id, that would work. If the files' extensions keep it from matching via Import with Matching, then why not add the extension to the name in the database, via a calculation field? Are the extensions all the same?

It would also be possible to parse the name without the extension later, via Left, as you say. If FileName is the imported name with an extension, the below will reliably remove the extension (even if there's other periods in the name).

Let ( [

len = Length (FileName);

pos = Position ( FileName ; "." ; len ; -1 )

];

Left ( FileName; pos - 1 )

)

Another idea, which may or may not work, would be to run a "PDF to text" converter on the image files. I have an old command line tool, pdftotext, which can do this, which can be found here (near bottom of page). There are other applcations that can do this, but I don't know if they can be run via AppleScript (this one can)

http://www.bluem.net/en/mac/others/

That would work if there is information IN the scanned file, and you can reliably get it out. But that would involve a more extensive AppleScript routine, which would then assign the file to a field belonging to that Transaction ID.

Another idea would be to create a folder structure for your existing Transaction IDs, that is, a folder for every one of them, with whatever additional structure required (whatever other IDs for enclosing folders). This is also an AppleScript job, but easier and faster than you think. Using the command line tool, mkdir, via AppleScript, you can create thousands of folders in a very short time. The idea being that you would scan into the correct folder (somehow). It is difficult to say what that "somehow" means, as it implies knowing what your scanner can do.

Another alternative would be to get the Troi File plug-in, which has a feature to enable "drag-and-drop" in a container field, or any area of a FileMaker layout.