Problem with table extraction

[IrinaOganezova]

The lines have different length, the extraction of pdf with following parameters

{"vertical_strategy": "lines",
"horizontal_strategy": "text",
"intersection_tolerance": 15,
"snap_tolerance": 3,}

results is

0 None 24
1 None

Please help with solution

[IrinaOganezova]

https://www.dgo.gov.pt/execucaoorcamental/SintesedaExecucaoOrcamentalMensal/2006/Janeiro/0106-bol.pdf

page 23

[jsvine]

Hi @IrinaOganezova, and thanks for providing the PDF. Parsing this in an automated way might be tricky (though possible, if you programmatically identify the location of the columns and headers), but a more manual approach like this could work:

table_settings = {
    "horizontal_strategy": "text",
    "vertical_strategy": "explicit",
    "explicit_vertical_lines": [ 70, 280, 325, 370, 410, 450, 488 ]
}

That will identify a table that contains your data — but which you might also want to clean up by first using page.crop(...):

[IrinaOganezova]

Thank you very much for help !