Question

Feature Request: Find index of a value from a multi-value token from a Zone OCR

Ideas Quick Fields Capture

Updated August 2, 2016

asked on March 20, 2014

Sometimes when using Quick Fields I know the general area for a ZoneOCR process but might have more information being captured than I need. I was hoping for a way to get the value of the index for a select phrase or word so I can use that to find the indices I need.

Say I have 6 lines coming in from the Zone, but the third line has the word, "Address is:" in it at the beginning of that line. And the next line will always be the start of the address I need to capture. But another time I get from the ZoneOCR 7 lines. This time the "Address is:" phrase is on the second line. The third line this time will be the first line I need to use to capture the data.

Having a function that can return the index of a phrase or word would be useful in identifying these changes of input quickly and plan accordingly.

0 0

Answer

SELECTED ANSWER

replied on March 20, 2014

You don't need a multi-value token in that case. If you read the text as a multi-line value, you can then specify a pattern matching that looks for "Address is:" ignores the rest of the line and gets the next one. Something like Address is[^\n]+\n+([^\n]+)

1 0

replied on March 21, 2014

Thanks Miruna,

I was not saying I cannot get around the lack of this feature, but rather, I wish we had this feature so if I have subsequent pieces of information the get from that original token, it becomes easier since I have the actual index.

I do have ways around this but it is rather nice when things are simplified and consistent so it takes less time to configure.

0 0

replied on March 21, 2014

Sure, expanding the token functions is on our list of features. But in this use case, reading the whole region in one value is the easier way, not the workaround since it allows you to get the value with just Zone OCR and Pattern Matching without any other token modifications or index calculations.

0 0

replied on August 2, 2016

I wish you had post some sample like when we have multi line zone and we are searching for example Remit to Address: ABC LLC.

We need to extract ABC LLC, and we do not know the line number. I Appreciate if your sample takes into consideration that only word Remit should be the start and the "to" and "Address" might be changed or not readable.

0 0

Replies

replied on March 20, 2014

Hi Kenneth,

How're you doing?

If I understand correctly, this is something that can be done by specifying the index of the match. Using a pattern match address, you could then grab the the subsequent lines...

%(token#[-1]#) for the last hit and %(token#[2]#) for the second.

Could you post the sample OCR'd text for me to play with?

-Ben

0 0

You are not allowed to follow up in this post.

Question

Question

Feature Request: Find index of a value from a multi-value token from a Zone OCR

Answer

Replies

Sign in to reply to this post.