>Hi José,
>
>>Could you expand a little more. I think it's the first time I read about this and I'd like to learn more.
>
>For starters, here's a little sample program that shows extreme speed differences between COLLATE and MACHINE. Run the program once with GENERAL and once with MACHINE to see what I mean.
>

>set collate to "general"
>set talk off
>set Safety off
>Sys(3054,11)
>
>create dbf dbf_a ( fielda1 N(2), fielda2 C(10) )
>for ll = 1 to 1000
>  append blank
>endfor
>replace all fielda1 with 99
>replace all fielda2 with "lol"
>goto 300
>replace fielda2 with "find this!"
>index on fielda1 tag fielda1
>index on fielda2 tag fielda2
>sele 0
>create dbf dbf_b ( fieldb1 N(2), fieldb2 C(10) )
>for ll = 1 to 1000
>  append blank
>endfor
>replace all fieldb1 with 99
>replace all fieldb2 with "xlol"
>goto 800
>replace fieldb2 with "find this!"
>index on fieldb1 tag fieldb1
>index on fieldb2 tag fieldb2
>set talk on
>select fieldb1, fieldb2 from dbf_a, dbf_b where fielda1 = fieldb1 and fielda2 = fieldb2

>Christof

Hi Christof!

I guess you are right. I have never thought that collate sequence cab affects the speed as much as I experienced in your test.
I will use Machine only. And will make some index tag to sort alphabetically with sys(15,...).
Do you have any additional advise?

BB

P.S.: After all, who dare to use collate sequences other than machine?