index problem

[Jeff Harrison]

----- Original Message ----

> From: Richard Kreiss <rkreiss at gccconsulting.net>
> To: filepro-list at lists.celestial.comf
> Sent: Sun, February 20, 2011 12:15:09 PM
> Subject: index problem
> 
> A client has been having an index problem after a status field value  
> is changed.
> 
> The record is accessed from a browse lookup by getting  the record  
> number and then doing a
> 
> 989  -------    -   -   -   -   -   -   -   -   -    -   -   -   -   -   -   -
> sho_rec? If:                                                               
>           ?
>         Then: lookup -  r=rn   -npw                                      
>            ?
> 990  -------   -   -   -   -   -    -   -   -   -   -   -   -   -   -    -   -
>         ? If: locked(-)                                                  
>           ?
>         Then:  MSGBOX "Someone Else is Updating this Patient  
> Record";GOTO  brw_new?
> 991  -------   -   -   -   -   -    -   -   -   -   -   -   -   -   -    -   -
>         ? If:                                                             
>           ?
>          Then: display(@sn);start_call=@tm                               
>           ?
> 992  -------   -    -   -   -   -   -   -   -   -   -    -   -   -   -   -   -
>         ?  If:                                                            
>            ?
>         Then: END                                                         
>           ?
> 
> If a sales is  generated, a processing table is called to complete the  
> operation.  Just before the processing table ends, the fields  
> status_sales, field 25, is changed to a 6.  The called table ends  
> returning the user back to the calling program and the original browse  
> lookup with the cursor on the next name.
> 
> Lately, but not always, they  get a message to rebuild index O which  
> has the primary key the status  field.
> 
> All updates to the record are done using @key  functions.
> 
> Would a write at the end of the called table avoid this  problem?
> 
> Richard Kreiss
> GCC  Consulting


It is difficult/impossible to tell why your index is getting corrupted without 
seeing all of the pertinent code.  Are you locking the record before the status 
is updated?  Are your doing a write or close to that protected lookup when you 
are done?   When doing something like you describe, I would do the protected 
lookup, change the data, then close the file.  After returning from the called 
process I would have it re-execute the original browse based on the new data.

Good Luck.
Jeff Harrison
jeffaharrison at yahoo.com
Author of JHExport and JHImport