I am having difficulty with tables in .pdf documents. First, ABBYY needs to scan a table and figure out how many columns are in it. With this public-domain document, it scans the top row of a table, and some of the cells framed by blanks are corrupted: It reads fails to see the blank cell and puts fewer cells in the output document. This usually happens on the top row.
When I go to fix table rows with fewer than 11 cells, it takes longer and longer as the issue if further down the document, and you need a feature to start at a user-specified page.
The other difficulty on some other documents is when on other table documents it splits a table down the middle as two tables when I need it to read as one. It needs an option, all tables on the page are one table.