sub extractor not working the same as normal pattern

Actually, your observations are 100% accurate. Sub-extractor patterns only match once. The intended effect is not "only make it match once", but rather, "look until you find this data, then save it". We would like to implement something to fix this issue, but for now, the solution is something like this:

1. Main extractor: as you already have it.
2. Call a script on that main extractor, "after each pattern application".
3. Make that script do the following:
   
   import com.screenscraper.common.*;

   text = dataRecord.get("DATARECORD");
   DataSet myDataSet = scrapeableFile.extractData(text, "Give a new pattern name here");
   

4. Create a new "main" extractor pattern, and put your original sub-extractor text in it. Name the pattern the same thing as the pattern name used in the script you just made.
5. Make this new "main" pattern only a manual-execution pattern, with the advanced tab of the new pattern.

Now you should be grabbing that info, but you'll have to process it in a script... Use the notes found in the API on the DataSet.extractData method. It shows you how to effectively loop through the results.

This has become a common dilemma with a few site structures, and we'd like to implement something to make it easier.

Hope that helps.
Tim