🚀 go-pugleaf

Hi Furno,

AFAIK, you have two alternative solutions:

1) you can pack the databases each time before rebuilding
   the index files. That way, the '*'-marked records will
   be deleted and cause no further trouble.
2) attach a filter expression to the index expressions to
   filter deleted records, e.g. "DELETED()". This way the
   index won't care for deleted records, which would also
   speed up searches a bit.
hope this helps,
Dirk

Furno Gianluca wrote:
> 
> Hi boys,
> 
> I'm using DBF DBase IV compatible tables with .MDX type indexes in my
> application
> 
> I have a problem using the DbiRegenIndexes function.
> 
> The sequence that unslash the problem is the following :
> 
> - I insert a lot of records ( The indexes are well updated )
> - I delete some records from the table ( I can see the '*' mark at the start
> of the record wich means that the record was correclty deleted in the DBF
> file)
> - I launch the DbiRegenIndexes function ( to simulate a lost of sync between
> Table <-> Index )
> 
> ----- Now i observe that inside the rebuilded index still contains the
> deleted KEYS ( Why?) ----
> 
> Now, at this point , If i try to insert the same KEY of a deleted KEY, the
> system throw an exception ( DUPLICATE UNIQUE KEY EXCEPTION OR SIMILAR )
> 
> Do i perform some wrong operations ?
> What is the right operation sequence to Delete and Re-Insert records with
> the same INDEXED KEY and how i can perform a safe index rebuiding ?
> 
> Please help me ....
> 
> Thanks,
> Furno Gianluca
> 
> info@microshadow.com

-- 
_________________________________________________
Dirk Schwammkrug
Computer Scientist R & D
Chemspeed Ltd, CH-Augst, http:\\www.chemspeed.com
Phone: (+41)61-8169-640  Fax: [..]-509