Bundle (collection) case sensitive

How To
,
1

Hello all,

Trying to figure out why the flow would break and finally figured out that the bundle is case sensitive to the text, eg “Description:”, “Unit:”, “Quantity:” etc.

I would not have expected this as I assumed the text that follows matters more whether they are number/text etc. However, it is this case sensitive issue that causes the problem when I parse documents and the output may say “Description” or “description” and this seems to impact the mapping in Google Sheets.

As you can see from the screenshots, when Google Sheets is happy the text boxes are a filled colour, but should the word change into lower case then it is an unfilled colour and nothing is passed from the sheet.

Any suggestions what I can add to make this agnostic to the letter case from the parsed documents?

lowercase
lowercase570×788 54 KB

Uppercase
Uppercase566×798 57 KB

success
success864×1288 45.2 KB

2

Ok managed to get support to solve this so posting it here in case someone else faces the same issue:

  1. Mapping Arrays points out that: " All parameters are case sensitive. Even though in this particular example the item’s label differs from its raw name only in capitalization, it is necessary to use the raw name, which is all lowercase value in contrast to the label Value" However the help section does not discuss on how to deal with instances like the one I mentioned: when a parameter varies after being parsed.

  2. The screenshot below shows what needs to be done and I was told to reference this guide if empty

    Screenshot 2024-12-13 at 5.12.17 PM
    Screenshot 2024-12-13 at 5.12.17 PM870×1128 50.2 KB

2 Likes
3

Hi @AlphaTee,

Thanks for getting to the bottom of this, I’m sure it will help many people out in the future!

Does this vary because the capitalization on the source sheet keeps changing?
Or do you know what was causing this?

1 Like
4

Welcome. In my specific case it was PDF.co parsing invoices. The data had the same structure but the parameters for one document (out of 3) had this variation.

I would assume it’s how the ocr at PDF.co end generates it even though there seemed to be no obvious indication from the documents as all of them had the same column headers and the same text (eg “Description”).

I can see this happening even if an email is parsed so I am glad they helped to sort it out!

5

I see. Good to know, thanks again.