My suggestion: Do a quick experiment. I had a similar question a year
or so ago. I wrote some code a year or so that generated random blobs
of varying sizes and tossed them into a SQLite DB and onto files on
the file system (Mac OS).
There are some complicating variables, such as our application
framework and some extra resource marshaling that might affect
performance one way or another, but our results were that for blobs
averaging <20-30K, SQLite was faster; for larger data sets, the file
system was faster.
One thing to consider is how the presence of said blobs might impact
performance of the database in general.
-Eric
On 21 Feb 2007, at 13:26, Brett Keating wrote:
Hi,
I'm curious about what the effect of having a blob in the database may
be on performance. I have two design options: 1) put a small image
file
(15-30kbyte) into the database as a blob, and 2) store the image in a
separate file on disk and hold the filename in the database. My table
has around 20 rows in it, about half are strings/smaller blobs and
half
are integers.
Option number one, for various reasons, is far more elegant and simple
in terms of its impact on the rest of the code. However, I am
concerned
that holding such large amounts of data per record might impact
performance. I could be worried about nothing though, which is why I'm
writing to this list :).
Any advice or input?
Thanks,
Brett
-----------------------------------------------------------------------------
To unsubscribe, send email to [EMAIL PROTECTED]
-----------------------------------------------------------------------------
