I would look towards how PostGis handles the Tiger census data for guidance. It's a similar, massive data set.
Greg Haase On Sep 20, 2013 9:47 AM, "Jeff Janes" <jeff.ja...@gmail.com> wrote: > On Thu, Sep 19, 2013 at 12:02 AM, Dave Potts <dave.po...@pinan.co.uk>wrote: > >> Hi List >> >> I am looking for some general advice about the best was of splitting a >> large data table,I have 2 different choices, partitioning or different >> schemas. >> > > > I don't think there is much of a choice there. If you put them in > different schemas, then you are inherently partitioning the data. It just > a question of how you name your partitions, which is more of a naming issue > than a performance issue. > > >> >> The data table refers to the number of houses that can be include in a >> city, as such there are large number of records. >> >> >> I am wondering if decided to partition the table if the update >> speed/access might be faster that just declaring a different schema per >> city. >> > > If you partition based on city, then there should be no meaningful > difference. If you partition based on something else, you would have to > describe what it is partitioned on, and what your access patterns are like. > > Cheers, > > Jeff >
