Hello,
Unfortunately PDFs don't have a concept of a "table". Every piece of information (line, word, sometimes every character) is a separate object, unrelated to others. In Ivy we call them "tokens" and Left() / Right() functions are simply moving to tokens on the left or on the right of the current one. Therefore, this behavior is correct.
Ivy has a few algorithms that can convert tokens into real tables. It's based on statistical analysis so results may vary, depending on complexity and "cleanliness" of the table. Please test it by selecting one of header tokens and clicking "Table" button in the toolbox. There are various options you can try to improve extraction, like using the border lines, amount of whitespace after the table, and so on. Please refer to "Table extraction" section in the Ivy documentation for more information, examples and tutorials.
Regards,
Andrew
