Lars B. Dybdahl wrote:

> What about this:
>
> Instead of putting large texts in table A, you add an integer that points to
> table B, which contains the texts. This way, you can reduce the number of
> texts in the system, saving space.

Since large blobs are stored off page, and blob pages aren't in the
table's pointer pages, you get that effect. What I was trying to
suggest was a way to discover that two documents - Word, PDF, or
text - were identical.

>
>
> I have such a system in one of my databases for log texts. Logs tend to
> repeat the same text again and again, and are normally very large in size.
>

I guess if you could recognize duplicate entries, you could have a
single copy of the text with many rows in the other table pointing
at it and keep only the message specific information in the referencing
table.


Regards,


Ann