A novice question (hopefully in the correct list):
I have installed PostgreSQL v6.5.3 on a Mandrake (RH 6.0) Linux system.
Regression tests do the normal thing....
Created a database "mydb", with a table "table1", comprising:
name type lgth null?
p_name bpchar 25 NOT NULL
s_region polygon
p_type bpchar 1
I'm trying to get data into this table, using the insert command to get
a feel for it, and the copy command for bulk insertions. (Longer term
goal is to have some code to load data from GIS packages into postgresql
for management/query systems)
insert into table1 ( p_name ) values ('poly1');
This works fine, the table has a single record, with NULL's in the other
two attrs.
insert into table1 ( p_name, p_type ) values ('poly1', 's');
This works, inserting a second record with only the polygon attr as
NULL.
insert into table1 values ('poly1',((0,0),(0,1),(1,0)),'s');
This fails with the msg:
ERROR: parser: parser error at or near ","
As I read the docs on polygon types, this should work. I can't find a
reference as to whether the polygon definition needs to repeat the last
coord pair to properly close the polygon or not. I have tried the insert
command with spaces between parentheses & values, and without commas
between polygon coords.
I need some help in getting carification of the polygon (& if possible,
other geometric datatypes) text formats for using with the insert/copy
SQL commands.
Related to this, given that the main purpose of the database will be
point-in-poly queries, some info on the relative merits of closed paths
vs polygons would also be appreciated (relative space, performance,
etc). I'm assuming (out of ignorance) that while they are functionally
equivalent, I'm better of using the polygon datatype in this context.
Any advice/suggestions gratefully accepted....
Thanks,
Brent Wood