Hi again, Naomi. Is it just my imagination, or haven't we talked
here on UT before? :^) :^)


>Just wanted to point out, that this is an old application converted and it uses old style
>of indexes INDEX ON somefield TO someindex.idx instead of index on somefield TAG Something.

I learned VFP by reading the VFP 6.0 book by Hentzenwerke, so some
of the syntax I use is the older style. Lately, though, I've been
using VFP's built-in help facility, so things are looking brighter.
Another thing I've found useful is to lurk here in the VFP groups,
listen to you guys define problems, and then look up every command
I have no clue about. After doing that, I see if I can find the
solution, and then I continue to read the thread until a solution
is provided, looking to see if I came up with the same solution
(or something there about). Sometimes, I actually get it right :^).

>You probably would not encounter a problem at all using structured indexes.
>
>Also I guess my record is out of range problem was mostly caused by bad indexes.

I was leaning in that direction, too, especially when I read the
chapter on row-column style indexing, knowing that if one little
thing was out of place...WHAM! :^)

> I did switch to array solution as a cleaner alternative and I also re-indexed my tables.

Kewl.

>I had this problem with local version of data and don't have a problem with the old code in production,
>so I really think the problem is related with the bad indexes...

Same here; however, I'm wondering how the indexes got thrashed to begin with?
I'm pretty sure I remember reading in the VFP 6 book that, from time to time, indexes
need to be rebuilt. It also mentioned that we should consider index corruption well
before we consider table corruption.

>BTW, I found select sum(..) from sometable where somecondition into array laArr doesn't create an array
>if the condition is not met in VFP8. Is the behavior different in VFP9?

Hmmm. Don't know. Since I'm not too busy today, I'll see if I can write a test program and find out.
Either that, or I'll sick one of my interns on the problem as an exercise :^). Hey, they might
as well learn what real-world code hackin' is all about :^).

Either way, it should be easy to figure out. Just run the select with bogus criteria, never
matching a record, and then see if the array exists. If it is created in 9, you'd think that
all the array elements would be NULL. (Shrug.)

>so I just put cRate = evl(laArr[1],0) (don't ask me about cRate for the numerical value, I didn't create such names)

Kewl. If it returns a value, calculate, and if it returns zero, then on to the next record.

Anywho, good luck on getting it working.

Regards,

Randall