[postgis-users] "Linux" geocoder script ?

Paragon Corporation lr at pcorp.us
Sat Mar 26 16:05:48 PDT 2011


> (One pendantic note -- "sh" or "bash" are *Unix* centric, not linux --
they are the standard on Freebsd, etc).
Yap you already said that more than once.

We changed the loader name to sh and also fixed some obvious typos we have.
We did have someone on linux test our older ones and he was able to get it
to work for him
with some minor typo corrections which we had integrated, but when we went
on to get it to work with 2010 data we accidentally introduced more typos.

So I really think the main issue is we didn't test it on Unix/Linux.

> So check out the function - loader_load_staged_data in tiger folder   and
> you'll see what I mean

> I think the loader function can hide some of that. Also, there is no way
to mix batch and sh scripts, you might as well keep them totally separate...

I disagree with your statement.  
The bulk of the logic is in the loader_lookup tables which as Steve pointed
out ARE going to have to change from year to year. That was what took us the
most time to do and when confronted with keeping the names of the columns
that census decided to suffix vs.
Having the latest year columns not have the suffix we opted for what we
thought was the easier to maintain which was -- have a plpgsql function
(RESUSABLE on all platforms), that would correct the issue so we wouldn't
have to change much of the geocoder from
Year to year.  That did require having a staging phase.

 I think it just didn't work out of the box for Linux because we didn't
test on Linux, because we were getting PAID to make it work on WINDOWS.
Really that boils down to it.

To us all the work we do on PostGIS that we contribute back is a side
product of our consulting work so what we contribute is really driven by
what we get paid for.  

Sorry to bring the whole money thing, but I feel people tend to lose sight
of that, that most of the PostGIS development team contributes back stuff
they were getting paid to work on and also work  in general that has no
strategic value
to be kept behind closed doors.

> I am glad those scripts work too.  The Unix / sh scripts though arent
anywhere near production ready.  Anywhere you see something like "%VAR%",
that is a batch file syntax that will throw an error in sh.  So,
create_gecoder.sh just breaks (I have a patch I will send later).
Great we love patches  :) 


postgis-users mailing list
postgis-users at postgis.refractions.net

More information about the postgis-users mailing list