🚀 go-pugleaf

If you are using the ODBC driver, you might want to try the native msaccess
driver.

My .02

Kevin

Alan Buckle wrote in message <01bdc84f$8c9e0620$01000a0a@pentium1>...
>Hi there
>
>I am using Delphi 3.02  (BDE 4.51) as a front end to an MSAccess 7.0 (95)
>databases.
>
>The application is being used at two sites, no WAN available, and I've
>written a program to
>reconcile the updates carried out at the 'remote' (or slave) site with the
>'master' database.
>(A sort of replication situation, but didn't want to use MSAccess
>replication for
>reasons which I won't go into here.)
>
>The program opens the two  databases and walks the slave looking for
>updated records which it
>then updates in the master (after some checks). So far so good.
>
>The database size is in excess of 100 MB and has approx 70,000 company
>records with ancillary 'child' records.
>
>However after processing in excess of 18,000 records (18408 to be exact,
>not counting 'child' records ) I get
>______________________________________________________
>Access Violation at Address 4E5FA1F6 in module 'IDDAO32.DLL'.
>Write of Address 0000 0000 .
>______________________________________________________
>
>while running under NT4
>
>or
>______________________________________________________
>Access Violation at Address 4E5FA1F6 in module 'IDDAO32.DLL'.
>Read of Address FFFF FFFF .
>______________________________________________________
>
>under Windows 95 on the same record.
>
>There is nothing apparently the matter with the data, starting the program
>a hundred records
>or so before the record will not generate the error (until it has processed
>another 18000 records.)
>
>The failure was remarkably consistent across different machines with
>different operating systems
>and differing amounts of memory.  Always failing at the same record.  This
>suggests to me that some sort of
>application 'quota' is being exceeded somewhere.
>
>Using the debugger and stepping through the program shows that the failure
>occurs on a 'Next'
>statement.
>
>The databases are not corrupt, the have been repaired and compacted and
>indeed regenerated.
>
>Restart the program half way through the database and it will again fail
>after 18000 records or so.
>(This varies slightly).
>
>I'm tearing my few remaining hairs out at this stage. Does anyone have any
>ideas ??
>
>I'm reluctant to phone Borland (InPrise??) for support as past experience
>has shown that this tends to be a waste of time
>and money, at least in the Netherlands.
>
>
>Many Thanks
>
>
>Alan Buckle
>
>APB Computing bv.
>
>email : buckle@euronet.nl
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>