Extracting Data into Columns

[chanpreet90]

I have a pdf which has data in tabular format and has 6 columns but the columns are not separated by boundaries so when I extract the data using pdfplumber, all the data comes in one column only and I want in separate columns.

How could I do that?

For your reference:
15/03/2021 RTGS-UTIBR52021031300662458-VIRENDER KUMAR 2,60,635.00 2,94,873.94Cr
11/03/2021 IMPS/P2A/107018040382/XXXXXXXXXX0980/trf 49,500.00 34,238.94Cr
11/03/2021 IMPS/P2A/107018771795/KINGDOMHOTELAND/trf 35,000.00 83,738.94Cr

[jsvine]

Hi @chanpreet90, unfortunately, it's difficult to help without access to the PDF. Could you attach it? I would also recommend trying the "text" strategies listed in the table-extraction settings documentation.

[chanpreet90]

here it is - https://we.tl/t-RuSjV2pXyh

[samkit-jain]

A word of caution ⚠️ @chanpreet90 Whatever redaction tool you have used hasn't done a good job. While the sensitive information like the account number is not visible, it is still very much copyable. Do Ctrl+A on the PDF and you'll see all the invisible text. Since it is a bank statement, request you disable the WeTransfer link to prevent any misuse. Going forward, please use sophisticated tools like https://github.com/JoshData/pdf-redactor to redact/delete any sensitive content from a PDF file.

[samkit-jain]

@chanpreet90 If "text" strategy doesn't help and there are no vertical separators, you can use the "explicit_vertical_lines" strategy. In your case, the extraction strategy as

{
    "vertical_strategy": "lines",
    "horizontal_strategy": "lines",
    "explicit_vertical_lines": [80, 360, 450, 560, 680]
}

works fine. It gives the following output

[samkit-jain]

If there is no \n, the output will be on the same line. \n represents a line break.

[chanpreet90]

Input -

Using extract_table() -

Converting into dataframe -

How to combine BALANCE and TRANSFER together like in the input image?

I hope I am able to make you understand my query.

[samkit-jain]

This is a different PDF from the one you shared earlier. Assuming you are using lines as the horizontal separator and getting this output, it most likely means that there are hidden lines in the statement. You can refer to #488 (comment) to know how to discard them.

[chanpreet90]

Yes, this PDF is different.
https://filetransfer.io/data-package/9pM0Upkg#link

[chanpreet90]

@chanpreet90 If "text" strategy doesn't help and there are no vertical separators, you can use the "explicit_vertical_lines" strategy. In your case, the extraction strategy as

{
    "vertical_strategy": "lines",
    "horizontal_strategy": "lines",
    "explicit_vertical_lines": [80, 360, 450, 560, 680]
}

works fine. It gives the following output

Why is there one blank row appearing when I join pages?