[postgis-users] Opinions for census type data base design

[Myunghwa Hwang]

Hi, list!
I am planning to make a web application which supports data extraction and
thematic mapping based on users' selection of what topic they want.
Currently, I am preparing database.
My problem is that I have quite a large of census items for four types of
administrative units.
Original source data are not structured in general gis data table, meaning
there is a large table that consists of three columns of variable id,
location id, value.
A typical database for web mapping is not usually in this structure. So, at
first I tried to pivot the large data table by variable id.
However currently, I have about 800, 800, 660, and 100 variables for each
type of administrative units.
Also, data tables after pivoting will have a large number of columns.
I am between two thoughts:
1) since my database will not be updated frequently, let's do pivoting, and
manage several tables with about 250 columns.
2) just keep the original structure for flexibility.
Which way do you is better? Any opinions are welcome.

from Myunghwa Hwang

-- 
Myunghwa Hwang
GeoDa Center
School of Geographical Sciences
Arizona State University
mhwang4 at gmail.com or Myunghwa.Hwang at asu.edu



-- 
Myunghwa Hwang
GeoDa Center
School of Geographical Sciences
Arizona State University
mhwang4 at gmail.com or Myunghwa.Hwang at asu.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20090210/7e58abc8/attachment.html>