Terry Vogelaar wrote: > David Vaughan heeft op zondag, 3 nov 2002 om 22:10 het volgende > geschreven: > >>> ... like David I really cannot imagine why Richard's suggestion about >>> using arrays would contribute anything to the speed, but I tried it >>> and it obviously worked; it reduced my (fastest) 101 seconds to 36! >> >> My problem was not in imagining it would be faster but in would be the >> code :-), given your particular problem. >> Would you mind showing your new solution, on or off list? > > Of course: > > put the number of lines of t into s > split t by return > repeat with f1 = 1 to s - 2 > put t[f1] into g1 > repeat with f2 = f1+1 to s - 1 > put t[f2] into g2 > repeat with f3 = f2+1 to s > put t[f3] into g3 > put g1 & "," & g2 & "," & g3 & return after outvar > end repeat > end repeat > end repeat > > You see, no big changes with the first script. It only uses an array > instead of the lines of a variable (25 secs faster) and outputs it to a > var instead of to a field (40 secs faster).
Excellent job, Terry! And thanks for posting that to the list -- this has been a valuable lesson in the benefits of employing Rev's unique syntax options. -- Richard Gaskin Fourth World Media Corporation Developer of WebMerge 2.0: Publish any database on any site ___________________________________________________________ [EMAIL PROTECTED] http://www.FourthWorld.com Tel: 323-225-3717 AIM: FourthWorldInc _______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution