Reading of table values in multi pages- How to avoid footer information

[KrishnaElapavul]

Hi
I have a scenario  where  I will read  values from a table which spread to  multiple pages in the pdf .
Decipher considering the footer values also  as a table data. how to avoid this  footer values ?
In the  given image decipher read yellow color data correctly , but red  underlined data also reading as table item which is not correct
Any advise to handle it ?

------------------------------
Krishna Elapavuluri RPA Solution Lead
TEchnology Consultant
DXC.technology
Asia/Kolkata
------------------------------

[lookman]

Hi Krishna.

How many documents have you used to train Decipher? I have worked with multi-page invoices and Decipher was able to learn the table boundaries after going through a handful of the documents.

jack

------------------------------
Jack Look
Sr Product Consultant
Blue Prism
------------------------------

[KrishnaElapavul]

Hi Jack,

I have processed 30+ documents of same vendor.

Thanks

Krishna

SKF

DXC Technology

T +91 99000 63042
Krishna-mohan.elapavuluri@dxc.com

DXC.com | Twitter* | Facebook* | LinkedIn* | Campaign Promotion*

Planned Leave: 14^th Oct,  2^nd &3^rd Nov

Company Holidays: 10^th Sep, 15^th Oct, 1^st Nov and 4^th Nov.

[lookman]

Hi Krishna.

Thirty docs should be enough. You could try using the misc param "BottomStop". I have not had to use it but see if it does the trick.

You'll find info on it here: Miscellaneous parameters (blueprism.com)

jack

------------------------------
Jack Look
Sr Product Consultant
Blue Prism
------------------------------

[KrishnaElapavul]

Thanks Jack,
looks this BottomStop is to stop lower lines
Still I tried, result is that , it stopping to read the other lines from next page onwards.


------------------------------
Krishna Elapavuluri RPA Solution Lead
TEchnology Consultant
DXC.technology
Asia/Kolkata
------------------------------

[lookman]

Hi Krishna.

I thought that might be the result but I wanted you to give it a try nonetheless.

The only thing I can think of is to keep pushing more representative documents through Decipher for training. I've had the same experience you have had but I was able to train Decipher to ignore them.

When you run your next doc, if the table is incorrectly mapped, delete the table and remap it yourself, taking care to only select the correct lines. It seems that deleting the table when it gets misread by Decipher helps in training.

jack

------------------------------
Jack Look
Sr Product Consultant
Blue Prism
------------------------------

[BenLyons]

Hi Krishna,

There's an option similar to BottomStop, but instead it skips any text area where it finds a specific word.

RowAreaExcludeValues

You can use this to say RowAreaExcludeValues=Tel in the Misc Params fields of the column where this might appear.

Note this can only be used for single words and not phrases.

Thanks

Ben

------------------------------
Ben Lyons
Product Consultant
Blue Prism
UK
------------------------------

[KrishnaElapavul]

Hi Jack/Ben

I  achieved with by adding 358 as exclude value . how ever this kind of number never expect in  my orders.
now i am able to skip footer information

------------------------------
Krishna Elapavuluri RPA Solution Lead
TEchnology Consultant
DXC.technology
Asia/Kolkata
------------------------------