[postgis-tickets] [PostGIS] #2288: Pre Standardize tiger input data

[PostGIS]

#2288: Pre Standardize tiger input data
----------------------------+-----------------------------------------------
 Reporter:  robe            |       Owner:  robe          
     Type:  enhancement     |      Status:  new           
 Priority:  medium          |   Milestone:  PostGIS Future
Component:  tiger geocoder  |     Version:  trunk         
 Keywords:                  |  
----------------------------+-----------------------------------------------
 As noted in ticket: #1118

 For issues such as those foreign abbreviations which depending on what it
 is and what county of the US you are in, e.g. Camino.  Such a thing may
 appear in full name, pre type post type etc.

 Prenormalizing the tiger data would solve this and be the best solution.
 I don't think I'll do this in 2.1 since it requires way too much forward
 thinking and testing which I simply don't have time for right now.

 but my proposed plan:

 1) Add an hstore column (call it norm_tags) to the featnames table

 Which would be essentially

 hstore(norm_addy)

 You ask why hstore?  Because
 1) I can index the pieces by key and I believe custom by value if needed
 2) I expect my norm_addy structure will be in flux for a while, so I don't
 want user tables affected by this and hstore seems like a stable enough
 structure to give what I need and shielding me from this issue.

 3) I'm deliberating on whether to do this only for problem cases or for
 all records.  I want to at least start of with just problem ones so that
 the scan list is shorter and after some benchmarking if it turns out more
 efficient to do for all, then move to that extreme.

-- 
Ticket URL: <http://trac.osgeo.org/postgis/ticket/2288>
PostGIS <http://trac.osgeo.org/postgis/>
The PostGIS Trac is used for bug, enhancement & task tracking, a user and developer wiki, and a view into the subversion code repository of PostGIS project.