Quick Fields pattern match works in testing, but not when session is run...

asked on July 10, 2015

I have a Zone OCR to capture all of the following:

       GST : Sub total: 14,073.24
       Tax: 703.66
       Invoice Total $14,776.90

and then Pattern Matching to return only the 703.66.

PROBLEM: The resulting field is blank when a session is run on a document, even though the pattern matching test returns the correct tax amount. I have researched this and others have had this problem, but I cannot see it resolved. One suggestion (found twice in ANSWERS) was that it may be finding a space where it isn't expecting it so try \s* every time there can be a space. So, I tried that in numerous forms and it did not resolve the problem. I tried (in the pattern matching configuration): Tax:\s*(.*); Tax:\s*(\s*.*); Tax:\s*(\s*.*\s*); Tax:\s*(\s*.*\s*)\s* all of which worked in the test area, but not when running the session on a document.

Output:

Can anyone help?

Thx, Connie Prendergast, Flagstaff County

0 0

replied on July 10, 2015

it looks as if you might have two spaces in that captured information, but i also dont see the tokens value that you have in the first screenshot showing in the secon screenshot. Are you sure you're evaluating the proper token?

0 0

View 6 previous replies

replied on July 10, 2015

Yes, the info is there...

0 0

replied on July 10, 2015 • Show version history

Yes but look at the token you defined at the "Edit Token" window

From what I see, you are actually using a different token, called %(Total, incl GST), which is not what you are showing us. So either we need to see what that token starts off as or you unintentionally assigned the wrong token to be used for that Pattern

0 0

replied on July 10, 2015

You're right, I had the wrong token there. Have corrected it, but still get blank results. There are five zones that I'm trying to collect here and they are all coming up blank. But thanks for pointing that out, Ken; I will double check them all.

RESULTS AFTER CORRECTING THIS TOKEN STILL COMES UP BLANK:

0 0

replied on July 10, 2015

Is it possible that Zone OCR and Pattern Matching running on different pages?

0 0

replied on July 10, 2015

I agree with Miruna, why not try and set that pattern to use the value of the input token when invalid, see what you get

0 0

replied on July 10, 2015 • Show version history

Checked the Pages settings. Both Zone OCR and Pattern Matching are set for Page One.

Ken, is this what you're suggesting (see next printscreen)? If so, I've tried it on both settings. I did notice that using this "mark fields...as invalid" it did not come up as Invalid even though it found nothing again. Tried numerous times. (Unless where it is marking it invalid is something I've not seen yet. I was looking for a red X on the document that was run through the session.)

0 0

replied on July 15, 2015

Did you set your OCR Zone process to multiline? It may not be reading the 2nd line where the tax info is at.

1 0

replied on July 15, 2015

I don't know how the originator's session is set up, but I can say that mine is set to read multi-line. If it wasn't then the sample would most likely not return the correct data.

0 0

replied on July 16, 2015

Wow, playing with these three Advanced Options (Single line, Create multi-value, Use existing text) is working. I now have four out of five of the fields being filled in with the correct answers! Thanks!

The only one not working now is the one that I did want to be a multi-value token. The Zone OCR is collecting everything shown below and the Pattern Match for that one is returning only the word "Description", but is not filling even that word into the field. (Below is a clip showing what I have in the pattern matching control window. I've been playing around and obviously the test isn't even turning up the right results right now. I did have the test working earlier. The target fields also have a maximum of 50 digits allowed, which might be part of the problem. The dashes it's picking up is from a thick border line that I can't get rid of.)

Description (for Subject line) : DESCRIPTION
--------- - _ _
CALCIUM CHLORIDE BRINE
WARD TRUCK CFR#
B/L# B/L# Haul TICKET
22253 25464
20696 25476
19726 25477
spray ticket: 18675
RR 153 North of HWY 13 Strome, AB
RR 153
RR 153

AHHH, testing more, removed the first asteric in DESCRIPTION\s*(.*) and now it is at least filling in the field with the word DESCRIPTION, which is the only word that I don't even need. I want the rest; each line as another line in the multi-line field!

0 0

Question

Question

Quick Fields pattern match works in testing, but not when session is run...

Replies

Sign in to reply to this post.