>Really? My question remains: Is it
possible to have a column where the columnorder is zero.
>
>My routine is not changing the column order, and is merely referencing it. It now stores it in a memory variable so that my error logging routine can see it. That memory variable is not changed any other way.
>
>Would it help if I said:
lnColumnOrder = this.columns(8).columnorder
??? I don't think that this would make any difference, would it?
>
>I'm not trying to be testy here, I'm just saying that I've already reduced this to a case where I KNOW that I have a value of zero for columnorder, and I'm trying to learn how this could possible occur.
>
>(And I must note that I cannot duplicate the problem, and the few users who get the problem get it very rarely, and cannot duplicate it once it happens.)
>
>>James,
James,
Actually I've seen errors without logical explanations. After trying to replicate them and spending some time I just gave up and put defensive code. In other words, you can always check for lnColIndex > 0 so it will eliminate these strange errors completely. I agree that mystery remains.
For instance, I have the following code failing in 2% of the cases
.oBizObj.requery()
saying something about unknown oBizObj. In order to eliminate this very strange problem I've changed this code to
if type('thisform.oBizObj.name') = 'C'
thisform.oBizObj.requery()
endif
Note, that I explicitly named thisform.oBizObj despite the fact I'm using with endwith construct. I've seen weird errors when . could not be resolved in some cases.
If it's not broken, fix it until it is.
My Blog