I agree with Graham that your document's overall structure is not conducive to data extraction.
If you're not already committed to the format you've submitted here, I'd strongly recommend reformatting it with a structure that is conducive to data extraction. You might, for example, create a 5-row by 1-column table for the 'header', and replace the nested 2nd table with a simple 8-row by 2-column table. In your 3rd table there should be separate rows for each item you want recorded in the Excel workbook (e.g. a separate row for each of attorney name, each line of the attorney address, phone, fax, email, etc., with every attorney having the same number of rows used the same way regardless of whether they're all filled in and each row being used for the same content (i.e. don't bunch the data up when there are missing lines). The same applies to litigants. If you don't do that, it will be well-nigh impossible to automate the data extraction in a way that ensures consistency of output.
As for your Proceedings table, I can only guess what the 'Proceedings text' column is supposed to contain; if the cells contain multiple paragraphs, that just adds to the data extraction woes...
Reply With Quote