On Wed, Jun 4, 2014 at 9:21 AM, Tom Lane wrote:
>
> If your input data is uniformly incompressible, that's not too surprising.
> pg_largeobject tuples hold BLCKSZ/4 bytes of data, plus some overhead, so
> the only way that 4 of them will fit on a page is if compression saves
> more than the overh
Jason Newton writes:
> I spent some time trying to get things to work as is, raising what limits I
> could to no avail. So I decided to upgrade to 9.3 and use large binary
> objects rather than making another file store due to a large convenience
> of keeping everything in database. I noticed t
Hi,
Some scope of setting: I use postgres to manage metadata about field tests
and simulations part of that involves HDF5 files. These hdf5 files are
generated both with field testing and simulations so there's going to be a
modest amount of them - in the 10k region eventually - an older database
Hi,
Some scope of setting: I use postgres to manage metadata about field tests
and simulations part of that involves HDF5 files. These hdf5 files are
generated both with field testing and simulations so there's going to be a
modest amount of them - in the 10k region eventually - an older database
