postgis/doc/faq.xml

363 lines
17 KiB
XML

<?xml version="1.0" encoding="UTF-8"?>
<chapter id="PostGIS_FAQ">
<title>PostGIS Frequently Asked Questions</title>
<qandaset>
<qandaentry id="legacy_faq">
<question>
<para>My applications and desktop tools worked with PostGIS 1.5,but they don't work with PostGIS 2.0. How do I fix this?</para>
</question>
<answer>
<para>A lot of deprecated functions were removed from the PostGIS code base in PostGIS 2.0. This has affected applications in addition to third-party tools such as
Geoserver, MapServer, QuantumGIS, and OpenJump to name a few. There are a couple of ways to resolve this. For the third-party apps, you can try to upgrade to the latest versions
of these which have many of these issues fixed. For your own code, you can change your code to not use the functions removed. Most of these functions are non ST_ aliases of ST_Union, ST_Length etc.
and as a last resort, install the whole of <varname>legacy.sql</varname> or just the
portions of <varname>legacy.sql</varname> you need.</para>
<para>The <varname>legacy.sql</varname> file is located in the same folder as postgis.sql. You can install this file after you have installed postgis.sql and spatial_ref_sys.sql
to get back all the 200 some-odd old functions we removed.</para>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>I'm running PostgreSQL 9.0 and I can no longer read/view geometries in OpenJump, Safe FME, and some other tools?</para>
</question>
<answer>
<para>In PostgreSQL 9.0+, the default encoding for bytea data has been changed to hex and older JDBC drivers still assume escape format. This has affected some applications
such as Java applications using older JDBC drivers or .NET applications that use the older npgsql driver
that expect the old behavior of ST_AsBinary. There are two approaches to getting this to work again.</para>
<para>You can upgrade your JDBC driver to the latest PostgreSQL 9.0 version which you can get from
<ulink url="http://jdbc.postgresql.org/download.html">http://jdbc.postgresql.org/download.html</ulink></para>
<para>If you are running a .NET app, you can use Npgsql 2.0.11 or higher which you can download from
<ulink url="http://pgfoundry.org/frs/?group_id=1000140">http://pgfoundry.org/frs/?group_id=1000140</ulink> and
as described on <ulink url="http://fxjr.blogspot.com/2010/11/npgsql-2011-released.html">Francisco Figueiredo's NpgSQL 2.0.11 released blog entry</ulink></para>
<para>If upgrading your PostgreSQL driver is not an option, then you can set the default back to the old behavior with the following change:</para>
<programlisting>ALTER DATABASE mypostgisdb SET bytea_output='escape';</programlisting>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>I tried to use PgAdmin to view my geometry column and it is blank, what gives?</para>
</question>
<answer>
<para>PgAdmin doesn't show anything for large geometries. The best ways to verify you do have data in your geometry columns are?</para>
<programlisting>-- this should return no records if all your geom fields are filled in
SELECT somefield FROM mytable WHERE geom IS NULL;</programlisting>
<programlisting>-- To tell just how large your geometry is do a query of the form
--which will tell you the most number of points you have in any of your geometry columns
SELECT MAX(ST_NPoints(geom)) FROM sometable;</programlisting>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>What kind of geometric objects can I store?</para>
</question>
<answer>
<para>You can store point, line, polygon, multipoint, multiline,
multipolygon, and geometrycollections. In PostGIS 2.0 and above you can also store TINS and Polyhedral Surfaces in the basic geometry type.
These are specified in the Open
GIS Well Known Text Format (with XYZ,XYM,XYZM extensions). There are three data types currently supported.
The standard OGC geometry data type which uses a planar coordinate system for measurement, the
geography data type which uses a geodetic coordinate system (not OGC, but you'll find a similar type in Microsoft SQL Server 2008+). Only WGS 84 long lat (SRID:4326) is supported
by the geography data type. The newest family member of the PostGIS spatial type family is raster for storing and analyzing raster data. Raster has its very own FAQ. Refer to <xref linkend="RT_FAQ"/>
and <xref linkend="RT_reference" /> for more details.</para>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>I'm all confused. Which data store should I use geometry or geography?</para>
</question>
<answer>
<para>Short Answer: geography is a new data type that supports long range distances measurements, but most computations on it are currently
slower than they are on geometry. If
you use geography -- you don't need to learn much about planar coordinate systems. Geography is generally best
if all you care about is measuring distances and lengths and you have data from all over the world.
Geometry data type is an older data type that has many more functions supporting it, enjoys greater support from third party tools,
and operations on it are generally faster -- sometimes as much as 10 fold faster for larger geometries.
Geometry is best if you are pretty comfortable with spatial reference systems or you are dealing with localized data
where all your data fits in a single <link linkend="spatial_ref_sys">spatial reference system (SRID)</link>, or you need to do a lot of spatial processing.
Note: It is fairly easy to do one-off conversions between the two types to gain the benefits of each.
Refer to <xref linkend="PostGIS_TypeFunctionMatrix" /> to see what is currently supported and what is not.
</para>
<para>
Long Answer: Refer to our more lengthy discussion in the <xref linkend="PostGIS_GeographyVSGeometry" /> and <link linkend="PostGIS_TypeFunctionMatrix">function type matrix</link>.
</para>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>I have more intense questions about geography, such as how big of a geographic region can I stuff in a geography column and
still get reasonable answers. Are there limitations such as poles, everything in the field must fit in a hemisphere (like SQL Server 2008 has), speed etc?</para>
</question>
<answer>
<para>Your questions are too deep and complex to be adequately answered in this section. Please refer to our
<xref linkend="PostGIS_Geography_AdvancedFAQ" />.</para>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>How do I insert a GIS object into the database?</para>
</question>
<answer>
<para>First, you need to create a table with a column of type
"geometry" or "geography" to hold your GIS data.
Storing geography type data is a little different than storing geometry. Refer
to <xref linkend="Geography_Basics" /> for details on storing geography. </para>
<para>
For geometry: Connect to your database with
<filename>psql</filename> and try the following SQL:</para>
<programlisting>CREATE TABLE gtest ( ID int4, NAME varchar(20) );
SELECT AddGeometryColumn('', 'gtest','geom',-1,'LINESTRING',2);</programlisting>
<para>If the geometry column addition fails, you probably have not
loaded the PostGIS functions and objects into this database. See the
<xref linkend="PGInstall" />.</para>
<para>Then, you can insert a geometry into the table using a SQL
insert statement. The GIS object itself is formatted using the OpenGIS
Consortium "well-known text" format:</para>
<programlisting>INSERT INTO gtest (ID, NAME, GEOM)
VALUES (
1,
'First Geometry',
ST_GeomFromText('LINESTRING(2 3,4 5,6 5,7 8)', -1)
);</programlisting>
<para>For more information about other GIS objects, see the <link
linkend="RefObject">object reference</link>.</para>
<para>To view your GIS data in the table:</para>
<programlisting>SELECT id, name, ST_AsText(geom) AS geom FROM gtest;</programlisting>
<para>The return value should look something like this:</para>
<programlisting> id | name | geom
----+----------------+-----------------------------
1 | First Geometry | LINESTRING(2 3,4 5,6 5,7 8)
(1 row)</programlisting>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>How do I construct a spatial query?</para>
</question>
<answer>
<para>The same way you construct any other database query, as an SQL
combination of return values, functions, and boolean tests.</para>
<para>For spatial queries, there are two issues that are important to
keep in mind while constructing your query: is there a spatial index
you can make use of; and, are you doing expensive calculations on a
large number of geometries.</para>
<para>In general, you will want to use the "intersects operator"
(&amp;&amp;) which tests whether the bounding boxes of features
intersect. The reason the &amp;&amp; operator is useful is because if
a spatial index is available to speed up the test, the &amp;&amp;
operator will make use of this. This can make queries much much
faster.</para>
<para>You will also make use of spatial functions, such as Distance(),
ST_Intersects(), ST_Contains() and ST_Within(), among others, to
narrow down the results of your search. Most spatial queries include
both an indexed test and a spatial function test. The index test
serves to limit the number of return tuples to only tuples that
<emphasis>might</emphasis> meet the condition of interest. The spatial
functions are then use to test the condition exactly.</para>
<programlisting>SELECT id, the_geom
FROM thetable
WHERE
ST_Contains(the_geom,'POLYGON((0 0, 0 10, 10 10, 10 0, 0 0))');</programlisting>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>How do I speed up spatial queries on large tables?</para>
</question>
<answer>
<para>Fast queries on large tables is the <emphasis>raison
d'etre</emphasis> of spatial databases (along with transaction
support) so having a good index is important.</para>
<para>To build a spatial index on a table with a
<varname>geometry</varname> column, use the "CREATE INDEX" function as
follows:</para>
<programlisting>CREATE INDEX [indexname] ON [tablename] USING GIST ( [geometrycolumn] );</programlisting>
<para>The "USING GIST" option tells the server to use a GiST
(Generalized Search Tree) index.</para>
<note>
<para>GiST indexes are assumed to be lossy. Lossy indexes uses a
proxy object (in the spatial case, a bounding box) for building the
index.</para>
</note>
<para>You should also ensure that the PostgreSQL query planner has
enough information about your index to make rational decisions about
when to use it. To do this, you have to "gather statistics" on your
geometry tables.</para>
<para>For PostgreSQL 8.0.x and greater, just run the <command>VACUUM
ANALYZE</command> command.</para>
<para>For PostgreSQL 7.4.x and below, run the <command>SELECT
UPDATE_GEOMETRY_STATS()</command> command.</para>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>Why aren't PostgreSQL R-Tree indexes supported?</para>
</question>
<answer>
<para>Early versions of PostGIS used the PostgreSQL R-Tree indexes.
However, PostgreSQL R-Trees have been completely discarded since
version 0.6, and spatial indexing is provided with an R-Tree-over-GiST
scheme.</para>
<para>Our tests have shown search speed for native R-Tree and GiST to
be comparable. Native PostgreSQL R-Trees have two limitations which
make them undesirable for use with GIS features (note that these
limitations are due to the current PostgreSQL native R-Tree
implementation, not the R-Tree concept in general):</para>
<itemizedlist>
<listitem>
<para>R-Tree indexes in PostgreSQL cannot handle features which
are larger than 8K in size. GiST indexes can, using the "lossy"
trick of substituting the bounding box for the feature
itself.</para>
</listitem>
<listitem>
<para>R-Tree indexes in PostgreSQL are not "null safe", so
building an index on a geometry column which contains null
geometries will fail.</para>
</listitem>
</itemizedlist>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>Why should I use the <varname>AddGeometryColumn()</varname>
function and all the other OpenGIS stuff?</para>
</question>
<answer>
<para>If you do not want to use the OpenGIS support functions, you do
not have to. Simply create tables as in older versions, defining your
geometry columns in the CREATE statement. All your geometries will
have SRIDs of -1, and the OpenGIS meta-data tables will
<emphasis>not</emphasis> be filled in properly. However, this will
cause most applications based on PostGIS to fail, and it is generally
suggested that you do use <varname>AddGeometryColumn()</varname> to
create geometry tables.</para>
<para>MapServer is one application which makes use of the
<varname>geometry_columns</varname> meta-data. Specifically, MapServer
can use the SRID of the geometry column to do on-the-fly reprojection
of features into the correct map projection.</para>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>What is the best way to find all objects within a radius of
another object?</para>
</question>
<answer>
<para>To use the database most efficiently, it is best to do radius
queries which combine the radius test with a bounding box test: the
bounding box test uses the spatial index, giving fast access to a
subset of data which the radius test is then applied to.</para>
<para>The <varname>ST_DWithin(geometry, geometry, distance)</varname>
function is a handy way of performing an indexed distance search. It
works by creating a search rectangle large enough to enclose the
distance radius, then performing an exact distance search on the
indexed subset of results.</para>
<para>For example, to find all objects with 100 meters of POINT(1000
1000) the following query would work well:</para>
<programlisting>SELECT * FROM geotable
WHERE ST_DWithin(geocolumn, 'POINT(1000 1000)', 100.0);</programlisting>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>How do I perform a coordinate reprojection as part of a
query?</para>
</question>
<answer>
<para>To perform a reprojection, both the source and destination
coordinate systems must be defined in the SPATIAL_REF_SYS table, and
the geometries being reprojected must already have an SRID set on
them. Once that is done, a reprojection is as simple as referring to
the desired destination SRID. The below projects a geometry to NAD 83 long lat.
The below will only work if the srid of the_geom is not -1 (not undefined spatial ref)</para>
<programlisting>SELECT ST_Transform(the_geom,4269) FROM geotable;</programlisting>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>I did an ST_AsEWKT and ST_AsText on my rather large geometry and it returned blank field. What gives?</para>
</question>
<answer>
<para>You are probably using PgAdmin or some other tool that doesn't output large text. If your geometry is big
enough, it will appear blank in these tools. Use PSQL if you really need to see it or output it in WKT.</para>
<programlisting>
--To check number of geometries are really blank
SELECT count(gid) FROM geotable WHERE the_geom IS NULL;</programlisting>
</answer>
</qandaentry>
<qandaentry>
<question>
<para>When I do an ST_Intersects, it says my two geometries don't intersect when I KNOW THEY DO. What gives?</para>
</question>
<answer>
<para>This generally happens in two common cases. Your geometry is invalid -- check <xref linkend="ST_IsValid" />
or you are assuming they intersect because ST_AsText truncates the numbers and you have lots of decimals after
it is not showing you.</para>
</answer>
</qandaentry>
</qandaset>
</chapter>