>I don't want to get into a Mohammed-like thread but, FYI, take the following line:
>
>QLN390030980 Ipswich: Sandyhill Lane, Roebeck Road E0057632
>
>"QL" introduces a bus stop rec., "N" means "new", 3900 is the county, 30980 is the stop #, the rest is the stop description
>
>QBN390030980 617249 242417 Ipswich
>
>"QB" is a continuation of a stop rec., the nos 617249 and 242417 are its lat/long, and Ipswich is the town
>
>>
>>QLN390030981 Ipswich: Sandyhill Lane, opp Roebeck Road E0057632
>>QBN390030981 617238 242410 Ipswich
>>Assuming lenghts are 20, 50, etc and all first lines of pair have ':' in it:
>
>Sore point! All the stop desc's USED to have this format, and the prog was so designed. Since this import file now has stops from other counties, one can't guarantee they have ":" in it. (some just have commas, others just one comma - can't guarantee what even more counties' data would be like).
>
>
>Jolly good idea all the same and I may just have a go at it.
>
>Thanks again and I appreicate the time and effort you've put into this.
>
>Terry
LOL M-like. No problem. I missed they were QL and QB but knowing them now my tendency for intermediate cursors is stronger. Instead of : QL and QB could be used as filters than (or Lat/lon fields containing alpha check). When you have time try once. It's really faster.
In my importer class I have no idea of what import file might look like. So first an Analyzer checks first 10 or so lines (getting into a cursor or fgets). It checks things like line lengths, tab, comma, quotes frequencies to decide on file type. Then presents the user all data in an intermediate cursor, file type 'sensed' and lets user to change type, mapping of fields etc. My initial version was using fgets and it needed a progress bar when dealing with over 20K records. Now current version still has the progress bar but unlike former all you might see is generally a flashing progress.
PS: You can check old version downloading FoxyClasses samples.
Cetin
Cetin