turbine-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Adam Sherman <a...@tritus.ca>
Subject Re: BLOBs (Binary Large Objects) & Torque/Turbine
Date Tue, 02 Oct 2001 14:09:53 GMT
On Tue, Oct 02, 2001 at 03:46:29PM +0200, Gunnar Rønning wrote:
> * Adam Sherman <adam@tritus.ca> wrote:
> |
> | Hmmm, interesting. I've built the prototype for this app in PHP using
> | Large Objects, and it has worked perfectly.
> | 
> | Note the advantage of using BLOBS: you access the files inside a
> | transaction.
> 
> The problem with PostgreSQL Large Objects is that they _must_ be handled
> inside a transaction. This has been an integration problem with Torque and
> underlying APIs, which has assumed another model.

This makes sense, however: does it mean that I cannot use BLOBs with
PostgreSQL & Torque/Turbine? Or is there a kludge?

> With PostgreSQL 7.2(which is about to go beta soon) we will have more
> standard BLOB support through the bytea datatype.
> 
> This is datatype is not as efficient has Large Objects, as it is part
> of the row and lacks interfaces to seek and read parts of the file. 

My problem is simple, the app is going to be used over local 100Base-T
by 75% of the users and will contain many 100MB plus files. This is
why I really need to use BLOBs, which the InputStream/OutputStream
that they support.

Thanks,

A.

-- 
Adam Sherman
President & Technology Architect
Tritus Consultant Group Inc.
+1.613.255.5164
http://www.tritus.ca

---------------------------------------------------------------------
To unsubscribe, e-mail: turbine-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: turbine-user-help@jakarta.apache.org


Mime
View raw message