Hello!

I have read in another thread how to handle a situation when the repair
function will not solve a corrupt IDX file.

Today we have had such a case, and running Dbsys repair fails with the
message: DBISAM Engine Error # 9217 Error reading from the table or backup
file 'xxxxx'.
As far as I know the only solution now would be to delete the IDX file and
recreate it with the proper indexes.
Apparently this works in our case!

So for the boneheads like me Tim, can you please spell it out, so that I can
understand why this is a bad solution?

Thanks
Petter Topp
ATC Data as

DBISAM Ver. 4.19

Petter,

<< So for the boneheads like me Tim, can you please spell it out, so that I
can understand why this is a bad solution? >>

It's a bad solution because you're erasing the index definitions along with
the data.  If you can send me the table in question, I can make sure that
the repair will handle the problem in the next build.

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim,

Just curious... is there a procedure/option anywhere that
will allow us to completely replace the current indices with
another set without first dropping the existing indices and
that works even if the current indices are corrupt or
missing/deleted? The purpose would be to handle corrupt
indices in one shot or to update the index structure without
worrying about what the existing indices are. Also, I seem
to recall that if I drop all the indices a primary index
will be created on the internal row id.


Sean



Tim Young [Elevate Software] wrote:
> Petter,
>
> << So for the boneheads like me Tim, can you please spell it out, so that I
> can understand why this is a bad solution? >>
>
> It's a bad solution because you're erasing the index definitions along with
> the data.  If you can send me the table in question, I can make sure that
> the repair will handle the problem in the next build.
>

Here it is.

Regards
Petter Topp
ATC Data as


"Tim Young [Elevate Software]" <timyoung@elevatesoft.com> skrev i melding
news:00ADEBDB-12E2-4401-A205-ED27B1897B71@news.elevatesoft.com...
> Petter,
>
> << So for the boneheads like me Tim, can you please spell it out, so that
> I
> can understand why this is a bad solution? >>
>
> It's a bad solution because you're erasing the index definitions along
> with
> the data.  If you can send me the table in question, I can make sure that
> the repair will handle the problem in the next build.
>
> --
> Tim Young
> Elevate Software
> www.elevatesoft.com
>
>

Sean,

<< Just curious... is there a procedure/option anywhere that will allow us
to completely replace the current indices with another set without first
dropping the existing indices and that works even if the current indices are
corrupt or missing/deleted? The purpose would be to handle corrupt indices
in one shot or to update the index structure without worrying about what the
existing indices are. Also, I seem to recall that if I drop all the indices
a primary index will be created on the internal row id. >>

If you copy a new .IDX over the old .IDX and then call RepairTable, it will
essentially do what you want.  However, again, this is not recommended.

--
Tim Young
Elevate Software
www.elevatesoft.com

Petter,

<< Here it is. >>

Thank you very much.  I will make sure that it is corrected in the next
build.

--
Tim Young
Elevate Software
www.elevatesoft.com