If I am reading this correctly, the collation has to be set before the dbe 
is loaded. If true, then in WAA (1.82 anyways) apps I do not have the option 
of changing this.

Is this correct?

I am using foxcdx with the 2x foxpro locking mode, but cannot get the 
collation to match what it was before so that functions such as descend() 
would return the same results.  And I can't use descend(expr, .t.) because 
this is not recognized by some of the tools I use (eg. Advantage ARC32).

Thanks in advance,

Les

In trying to find a solution, I discovered something interesting: setting 
the CHARSET to ASCII prior to building the index results in descend() 
working as expected, and subsequent updates to the dbfs with our usual 
charset set to ANSI seems to update the index properly.

Does this make sense?

Here is the environment:

- i converted our dbfs using a one-time standalone xbase app to foxpro dbfs 
with the 2x locking mode.

- everything is done using a waa app, using foxcdx and charset ANSI, using 
indices with comix compatibility mode

- when our waa app rebuilds the indices, descend() seems to do nothing 
resulting in an incorrect sort order. if i rebuild the same indices using 
ARC32, the sort order is once again correct

- if i set the charset to ASCII before rebuilding the indices in our waa 
app, the sort order is also correct.

- our waa app, running with charset ANSI, updates the dbfs and the sort 
order remains correct

Regards,

Les

"Les C. Cseh" <les.cseh@asapchecks.com> wrote in message 
news:3bc7823a$2b33f031$1cc364@news.alaska-software.com...
> If I am reading this correctly, the collation has to be set before the dbe 
> is loaded. If true, then in WAA (1.82 anyways) apps I do not have the 
> option of changing this.
>
> Is this correct?
>
> I am using foxcdx with the 2x foxpro locking mode, but cannot get the 
> collation to match what it was before so that functions such as descend() 
> would return the same results.  And I can't use descend(expr, .t.) because 
> this is not recognized by some of the tools I use (eg. Advantage ARC32).
>
> Thanks in advance,
>
> Les
>
>

KRAP, I stand corrected. Unless charset is ASCII, descend() does not appear 
to work with foxcdx/fox compatibility mode.

It looks like the only answer in my case is to rewrite to that I do not use 
descend() until I can go to VFP compatibility mode.

Les

"Les C. Cseh" <les.cseh@asapchecks.com> wrote in message 
news:5b70b534$60123e98$240f28@news.alaska-software.com...
> In trying to find a solution, I discovered something interesting: setting 
> the CHARSET to ASCII prior to building the index results in descend() 
> working as expected, and subsequent updates to the dbfs with our usual 
> charset set to ANSI seems to update the index properly.
>
> Does this make sense?
>
> Here is the environment:
>
> - i converted our dbfs using a one-time standalone xbase app to foxpro 
> dbfs with the 2x locking mode.
>
> - everything is done using a waa app, using foxcdx and charset ANSI, using 
> indices with comix compatibility mode
>
> - when our waa app rebuilds the indices, descend() seems to do nothing 
> resulting in an incorrect sort order. if i rebuild the same indices using 
> ARC32, the sort order is once again correct
>
> - if i set the charset to ASCII before rebuilding the indices in our waa 
> app, the sort order is also correct.
>
> - our waa app, running with charset ANSI, updates the dbfs and the sort 
> order remains correct
>
> Regards,
>
> Les
>
> "Les C. Cseh" <les.cseh@asapchecks.com> wrote in message 
> news:3bc7823a$2b33f031$1cc364@news.alaska-software.com...
>> If I am reading this correctly, the collation has to be set before the 
>> dbe is loaded. If true, then in WAA (1.82 anyways) apps I do not have the 
>> option of changing this.
>>
>> Is this correct?
>>
>> I am using foxcdx with the 2x foxpro locking mode, but cannot get the 
>> collation to match what it was before so that functions such as descend() 
>> would return the same results.  And I can't use descend(expr, .t.) 
>> because this is not recognized by some of the tools I use (eg. Advantage 
>> ARC32).
>>
>> Thanks in advance,
>>
>> Les
>>
>>
>
>