On Fri, Apr 16, 2010 at 12:40:21PM +0200, Szymon Guz wrote:
> I thought that the default fillfactor was much smaller (and haven't checked
> that now)... sorry for messing that up.
> But let's think of it from the other side: what do you want to do with that
> data? Maybe PostgreSQL with it's MVCC's overhead isn't the best solution for
> your needs.
I'm using this as part of a larger application. The data sets are one aspect
of it. The idea is the following:
An engineering application generates time-dependent data. One simulation
yields a very big text file in tabular format, with hundreds or thousands
of columns with output values (often more than Postgres' column limit),
one row per timestamp.
One such file is generated for each permutation of input values which
influence the run of a simulation.
This text file is imported into a database so we can perform very quick
lookups on the numbers so they can be quickly plotted in a graph.
The user can select any number of input permutations and graph the
values of any selected output values to view the effect of the variation
in input.
One can plot any variable against any other, so one join is made for
each variable that we want to plot; it joins the timestep values of the
variable on the X axis to those on the Y axis.
Regards,
Peter
--
http://sjamaan.ath.cx
--
"The process of preparing programs for a digital computer
is especially attractive, not only because it can be economically
and scientifically rewarding, but also because it can be an aesthetic
experience much like composing poetry or music."
-- Donald Knuth