Обсуждение: large object minimum storage

Do large objects have to be at least 16K, plus 16K for the index, even if
the object is much smaller?

I have a bunch of large objects of widely varying size.  Most are quite
small, but a few are above the 8K limit, so I thought I might just store
them all as large objects.  But this was taking up far too much space, and
I noticed that the xinv* and xinx* files were all 16K, which is much more
than most of them need.

Is this a configuration parameter?  I didn't see anything in the
documentation.  (I am currently using 6.3.2 on Redhat linux 5.2.)

I've pretty much decided to use arrays instead of blobs, and put the
"overflow" in a separate table for the few objects that exceed the limit,
but if anyone has any better ideas, I'd love to hear them.

Thanks

Ken

6.5 I think uses less space, but it still has to be a multiple of 8k
blocks.

> Do large objects have to be at least 16K, plus 16K for the index, even if
> the object is much smaller?
>
> I have a bunch of large objects of widely varying size.  Most are quite
> small, but a few are above the 8K limit, so I thought I might just store
> them all as large objects.  But this was taking up far too much space, and
> I noticed that the xinv* and xinx* files were all 16K, which is much more
> than most of them need.
>
> Is this a configuration parameter?  I didn't see anything in the
> documentation.  (I am currently using 6.3.2 on Redhat linux 5.2.)
>
> I've pretty much decided to use arrays instead of blobs, and put the
> "overflow" in a separate table for the few objects that exceed the limit,
> but if anyone has any better ideas, I'd love to hear them.
>
> Thanks
>
> Ken
>
>
>
>


--
  Bruce Momjian                        |  http://www.op.net/~candle
  maillist@candle.pha.pa.us            |  (610) 853-3000
  +  If your life is a hard drive,     |  830 Blythe Avenue
  +  Christ can be your backup.        |  Drexel Hill, Pennsylvania 19026

On Sat, 26 Jun 1999, Kenneth Been wrote:

# I've pretty much decided to use arrays instead of blobs, and put the
# "overflow" in a separate table for the few objects that exceed the
# limit, but if anyone has any better ideas, I'd love to hear them.

    The most common way I've seen (and done) is to put the stuff in a
table with an id, a sequence number, and the data you want to store.  I
store image data by base64 encoding it and storing the pieces in little
individual rows of about 76 characters each.  It works out pretty well
that way.

--
SA, beyond.com           My girlfriend asked me which one I like better.
pub  1024/3CAE01D5 1994/11/03 Dustin Sallings <dustin@spy.net>
|    Key fingerprint =  87 02 57 08 02 D0 DA D6  C8 0F 3E 65 51 98 D8 BE
L_______________________ I hope the answer won't upset her. ____________