RE: DATALINK (was: How can I store an externally generated PDF inafile or outq)

Thanks Keith, that was along the lines that I was thinking of, although I
was having difficulty conceptualizing a 16MB field that a BLOB could be!

The other difference I just thought of is that the same DATALINK could be
used by more than one record.  That couldn't be done with a BLOB.  Although
that starts to get into some data normalization issues I suppose.

Anyway, how does one choose between a BLOB and a DATALINK?  I already
described how I see this being used to pop up photos, drawings, and
signatures in a window via an F-key.  But the application that we will have
is to print reports (maybe produce PDFs?) that include these graphics as
well.

db

> -----Original Message-----
> From: midrange-l-bounces@xxxxxxxxxxxx / Keith Carpenter
> Sent: Friday, October 08, 2004 1:13 PM
>
> The main difference is where the data is stored.
>
> With a BLOB the data is considered stored in the row (record).  Actually
> it's probably some space object behind the scenes (there's a 16 Mb size
> limit). You can think of a BLOB as a very big data field.
>
> With a DATALINK the data is in the IFS or some remote system and it's
> referenced by URL instead.  Hence why it's called a link.  You
> can think of
> the field as containing the path to your stream file.
>
> A windows app could use ODBC to read your image db, then use the URL
> w/access token to open the image file from the IFS.
>
>
> Keith