Chuck Jungmann wrote:
Responding to both of your comments:
My client regularly acquires a database of property maps from a county.
The database is a ESRI Personal Database. An employee of my client has
been converting this database (and others like it) to MapX files for
running on MapXtreme (on Windows, as you observed). I understand it to
be a rather involved process.
On the subject of creating a shape file from the PGeo file: one of the
layer tables has almost 200,000 records. That's just for the parcel
outlines. I also would need to add the text labels and other outlines as
well. That would make for an unwieldy shape file. In fact, I think
that employee mentioned above actually does combine the shapes into 9
MapX layers that I have been combining for display. The combined MAP
files come to 221.3MB, and 351.6 MB for all layer files.
I am exploring the possibility of delivering the same web application
using open source software, including MapServer, on Linux. I was hoping
to use the database as delivered, because it would save time and
effort.
Although it looked like it should have been possible to display layers
from the PGeo file, I haven't had any luck. I have spent a lot of time
trying to figure out if I could apply some patch, looking for
insufficient buffers that I could lengthen, etc., but have pretty much
come to the conclusion that using PGeo files is perhaps only feasible on
MapServer for Windows, where there is native support for the MDB files.
I originally asked the question because
1. I didn't want to give up too soon if someone has had some success
using unixODBC and MDBTools, and
2. I wanted to know if making an exact copy of the PGeo tables in a
MySQL database would allow me to use PGeo to access the layers. This
seems feasible since both MDBTools and MySQL can be accessed through the
ODBC layer.
It would be nice if OGR wouldn't know the difference between MDB and
MySQL databases because I could then run a script (that I'm working on
now) that would run on Windows to make a restore file for MySQL to
import the data. I hope to finish the script today or tomorrow. If it
works as I hope, I'll report it here since it would answer the second
question.
Assuming success using MySQL, the first question is moot: MySQL would be
a much more robust solution. If I can make a turnkey conversion from
MDB to MySQL, it's really a no-brainer to abandon MDB. But I'm still
curious, and it would be nice to document somewhere online, if it's
possible to use PGeo files on a Linux-based MapServer.
Chuck,
Personally a little ogr2ogr script would convert the mdb layers into
shapefile layers, and then the same script could execute shptree on each
new shapefile. One of the (unannounced) results that I saw in the
recent benchmarking exercise at the FOSS4G event was how fast large
shapefiles (over 1,000,000 records) that are indexed properly are drawn
by MapServer.
So my advice is to keep it simple: use MS4W on windows, write a script
that can be executed each time your client hands you a lovely PGeo file,
and create the associated layers in a MapServer mapfile.
-jeff
--
Jeff McKenna
FOSS4G Consulting and Training Services
http://www.gatewaygeomatics.com/
_______________________________________________
mapserver-users mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/mapserver-users
