Thanks for the feedback so far. In terms of the layout for the new multiple instance data, the reason why it is output in a flat format is that otherwise every combination of pattern matches would need to be output in order to replicate the combined table layout. Since this could be very large when multiplied out by each multiple instance, we decided to keep the data as a flat output in the order that the data was found in the PDF files.
One way to reformat this this as column is to use the multiple instance data table as a source for a pivot table and then layout the results in the desired fashion for analysis. Any ideas would be appreciated.
