Subject Re: SEARCHING VERY LARGE FILE From Ronnie MacGregor <No_Sp@m.Thanks> Date Thu, 2 Sep 2021 19:47:36 +0100 Newsgroups dbase.getting-started In article <IGAYdN9nXHA.1636@ip-AC1E04A7>, invalid@invalid.invalid says... > > I prefer waiting for 1 second before requery() > > Agostinho is using LIKE '%whatever%' to select records from over 1 > million records in a .dbf file.  This is a hard task for SQL and takes > long enough without adding a delay between sending individual characters > of the search string to the SQL engine. Actually I use a similar approach for an intellegent search form which with a single entryfield will try to establish the type of data that is being searched for and query accordingly. For example three letters from a first name and three letters from a surname is usually enough to find a Patient, searching two fields, and coping with the search data presented the other way around. And a while array of other bits of data, DOB, PxID, etc. etc.. Anyway . . . . The logic of the delay in triggering a requery is that you reduce the number of queries being made, and allow the user to type three or four characters before a search is initiated. From memory, I think the delay the we settled on is around one second.