What influences the result of extracting a PDF table to Excel?

I have a PDF file with multiple tables, but when extracting, only some tables are obtained. Suspecting that AI might be confused when dealing with multiple tables, I separated one table into a single PDF file and used a table extraction API tool (WonderShare PDF). The result did not match the original Excel table.

I used the pdf-lib library and code to separate the table into a new PDF file. After extraction, the Excel result seems to be not entirely correct.

What factors could affect the results of extracting tables using the PDF API? I am attaching two PDF files and two result files when using WonderShare PDF tool and PDF-lib.