1 Reply Latest reply on Mar 14, 2013 6:03 AM by mikebeargie

    Auto-Complete Indexing Question




      We have a db with 150K records in it. These are people who take a specific training course and are usually associated with a Company. We have "auto-complete" set on the company field. The first attempt at entry into this field can take a while. We are wondering if we can speed it up by cleaning the existing data.


      Pretty safe bet that:


      Alliance Refactory
      Alliance Refortorties
      Alliance Refractories
      Alliance Refractories
      Alliance Refractories ltd


      are the same company. The exported sorted list reveals 7K names, a cursory examination suggests that this might be whittled down to 2K with modest cleanup.


      So the question is whether it is the number of records (150K) or the number of unique values (2K to 7K) that has the greatest impact on building of the Auto-Complete list. We are trying top avoid having a user managed list.




      "Maybe that's what life is... a wink of the eye and winking stars." --- Jack Kerouac

        • 1. Re: Auto-Complete Indexing Question

          Indexing will look at every record's value while building the index. So I'd imagine the number of records is causing it. IE, you're not indexing 7000 record values, you're indexing 150,000 record values which equates to a list of 7000 values.


          How are you storing the field's index? Is it set to store "all"? (vs. minimal or as-needed). It might be worth it to turn on all storage for this one field. It will increase your file size, but cut back on the initial index creation it seems to do the first time that field is used.


          Here's a filemaker article about indexes for reference: