Re: Performance suggestions welcomed

You don't say how you are doing this processing. But assuming you have written it in 4GL why not do it SQL as a SELF-JOIN. SELECT a.field,b.field from tabname a,tabname b where a.field = b.field An index on the field would make it run quite well. > Any ideas on how to increase performance on the scenario detailed below > ; > > Reference info : V7 online on SCO, > a table with 1 column of char(30), 50,000 records. > > Processing : read first record, compare this row with other 49,999, > then read next row , compare with other 49,999, > etc. etc. etc... > > Even ignoring the compare (string matching) part above, it takes a hell > of a long time, hours and hours and hours. Any tips like, read the lot > into memory, buy a CRAY, etc. etc. would be welcomed. > > reply piyal@piyalf.demon.co.uk > > Malcolm Weallans Online Database Consultancy Phone 01628-72154 Fax 01628-37463 CIX - onlinedbc