Hi Avina,
The setting looks correctly added.
Unfortunately you have a situation that doesn't have an obvious answer, so it will be a case of trying a few different ideas. The challenge you have is that every time you train a document with a technique that doesn't work, your training data will be a little further away from where you need it. This is why I regularly delete my training data during configuration, as this helps speed up the learning.
I can't see clearly, but it looks like your sample header is "CUST PART #:", when the text is only "CUST PART #" (without : ). Is this correct?
You could perhaps also try it without the sample header and maybe add fields for the other columns of the table. Even if they're not needed, it may help Decipher learn the layout. You can always add the flag "Non Exportable" and they won't be sent to Blue Prism.
If that doesn't work, you could try a couple of the sub-table misc parameters
http://bpdocs.blueprism.com/decipher/user-guide/misc-parameters.htm.
"SubTableCategory=On" may help show that this is part of a slightly different layout.
Ultimately with non-standard layouts like this, you will need to look through the available functions and see how they might help.
Thanks
------------------------------
Ben Lyons
Product Consultant
Blue Prism
UK
------------------------------
Ben Lyons
Principal Product Specialist - Decipher
SS&C Blue Prism
UK based