Hi, I am very new to linux.. please bare with me if I am asking stupid question.

I have a bunch of pdf that may not be written in format (randomly downloaded from google scholar). I am trying to check through their data source, so I used pdfgrep for the task.

It was successful most of the time until I found out that are some tables vertically placed. I attempt to rotate before grep-ing them
but it makes no difference.

then I tried

which actually served my purpose. It rendered some caption line in the vertical page but the vertical table is messed up, the numbers are in chaos.. and there are some numbers missing.

I would like to know if there is any more elegant way to complete the task?

The .pdf is here.

Thank you.

Hi fdq09eca,

The link to your PDF gives a preview only and requires to log in for the full PDF.
If you can (legally accepted) give me access to the full PDF, I'm willing to do some testing for you with a completely different tool.
I'm talking about Apache TIKA. Open source under Apache license. I used this heavily in forensics.
Of course yo can download (https://tika.apache.org/download.html) and try for yourself.
It is developed for the extraction of text and/or metadata from loads of filetypes, including PDF. It can output txt, html, xml, json.

This is an article from 1995. E.g. here the text layer has only the title. If the version at JSTOR was scanned from paper and then OCRed then the text you get from it is only as good as the OCR layer saved in the PDF. Rotating the pages with pdftk, mutool etc. won't change anything if the tables weren't OCRed properly in the first place.

1 members found this post helpful.