Extract table in PDF

(1) Share

Report

Posted on by JoaoCord

Dear Members,

I am new to Power Automate Desktop.

I want to extract this table from this pdf file and convert it to Excel, however I ran into a problem.

When I use the ExtractedPDFtables action, the table only has four columns, not the six that appear in the original PDF. The first two are not extracted.

Could you please explain why and how I could overcome this?

Thanks in advance for your assistance.

Livro2.pdf

Categories:

Power Automate Desktop

I have the same question (0)

All responses (6)

Answers (0)

Sort by

UshaJyothiKasibhotla 225 Moderator on at

Like
a
(0)

Report
Copy link

Link copied!

Hi @JoaoCord

Sorry for the delay...

I have seen it in unanswered questions.

Could you please send me the sample pdf so that I can give a try and get back to you.

Was this reply helpful? Yes No
JoaoCord 8 on at

Like
a
(0)

Report
Copy link

Link copied!

Hi,
I am sending herein attached the pdf file that I would like to get in a excel table.
The issue is that when i use "ExtractedPDFtables" action from power automate desktop, the table only has four columns, not the six that appear in the original PDF. The first two are not extracted.
Thanks in advance for your assistance.
Best regards.

Livro2.pdf

Was this reply helpful? Yes No
CU16071609-1 6,255 Moderator on at

Like
a
(0)

Report
Copy link

Link copied!

@JoaoCord
I think there might be an issue with the PDF table format, resulting in only four columns being displayed instead of six.
To troubleshoot, I recreated a table(Attached for your reference) with similar data in Excel, pasted it into a Word document, and then converted the document to a PDF file. Upon extracting data from the recreated PDF file, the correct data was displayed.
Please see the details of the data tables below for your reference, where I compare your original PDF datatable with the recreated PDF datatable.

So I suggested you to correct the table format in the pdf document else extract the data as text using the PAD action "Extract text from pdf" where it will extract the all data but we have to do some workaround to get it formatted.

Extracted data as textL

Thanks,
Deenu

Formattedtable.pdf

Was this reply helpful? Yes No
JoaoCord 8 on at

Like
a
(0)

Report
Copy link

Link copied!

Hello Deenuji,
Thanks for your kind help.
The problem is that I received the pdf file from a professional contact and i should work over it.
It is a document with multiple pages but in some pages the PAD flow doesn't recognize 6 columns but only 4.
If i have to do all the steps that you are proposing, besides the time consumption, it will lose all the power that PAD as to offer.
My goal is to transform the PDF data to excel automatically.
Thanks again. Joao

Was this reply helpful? Yes No
Agnius Bartninkas Most Valuable Professional on at

Like
a
(1)

Report
Copy link

Link copied!

The reason for this is likely because the rows in the first two columns are merged, while they are not merged in the other 4 columns. When the row structure is different, PAD does not recognize it as a part of the same table.

The likely only way to work around this would be to extract it as text via Extract text from PDF and then use some regular expressions and other text manipulation actions to retrieve the data you want. But it will still likely be a bit of a pain to do with this sort of a formatting, as it will likely be difficult to identify which rows in the first two columns should correspond to which rows in the other four columns.

Was this reply helpful? Yes No
CU16071609-1 6,255 Moderator on at

Like
a
(1)

Report
Copy link

Link copied!

@JoaoCord
As previously mentioned, the issue lies with the PDF table's formatting, as it can only identify four columns and not the remaining two.

This isn't a critique of PAD's capabilities; rather, it reflects the typical functioning of RPA algorithms. If the table is properly formatted and structured in the PDF, recognition becomes easier. However, if it isn't, resorting to the "Extract text from PDF" action, as mentioned earlier, may be necessary.

Nonetheless, this approach entails significant effort in formatting the table from plain text, where discerning blank columns and values becomes challenging.

Regrettably, this seems to be the only viable solution for your current scenarios.

Thanks,
Deenu

Was this reply helpful? Yes No