You are working with audio recognition data tables such that each table’s initial and final rows contain zeroes. These beginning and ending zeroes are irrelevant noise that must be removed. Here is an example of such a data table: 0 0 8 0 7 1 0 Your goal is to only remove these unnecessary starting and trailing zeroes. You must keep the zeroes in the middle of the data, which do not constitute noise. As well, the position of the starting and trailing zeroes differs per data table, so a good solution probably requires flow variables for flexibility (although there may be other solutions which do not involve flow variables).