hobu / crs Goto Github PK

I've been doing a little bit of prototyping with the EPSG DB in Python and I ran into this problem.

The EPSG SQL files are encoded by default in ISO-8859-1. SQLite assumes that text data should be encoded in Unicode format.

How do you test this out to see if everything is working properly?

$ sqlite epsg.sqlite
sqlite> SELECT * FROM epsg_unitofmeasure WHERE uom_code=1029;

which should return this line

1029|year|time|1040|31556925.445|1.0||International Union of Geological Sciences (IUGS) and International Union of Pure and Applied Chemistry (IUPAC). Pure Appl. Chem., Vol. 83, No. 5, pp. 1159–1162, 2011.|OGP|2013-05-28|2013.021|0

The 1159–1162 should be an en-dash which is U+2013. Otherwise, it should be some other odd character. (In Python, the SELECT causes an Exception to occur. SQLite's console utility seems to just fill in what character it thinks it is.)

The solution is the data file needs to be converted to UTF-8, using iconv. (In Windows, I used notepad to save as UTF-8, but you could install iconv port.)

~/epsg2$ file EPSG_v9_0.mdb_Data_PostgreSQL.sql 
EPSG_v9_0.mdb_Data_PostgreSQL.sql: ISO-8859 text, with CRLF line terminators
~/epsg2$ iconv -f ISO-8859-1 -t UTF-8 EPSG_v9_0.mdb_Data_PostgreSQL.sql -o  out_utf8.sql
micah@micah-VirtualBox:~/epsg2$ file out_utf8.sql 
out_utf8.sql: UTF-8 Unicode text, with CRLF line terminators

IMO iconv should be added as a step to the build.sh.

I tested both the issue and the solution on Ubuntu 16.04 and Windows 7.