Download Lecture 2

Spatial Databases: Lecture 2
DT249, DT211,DT228 Semester 2
2016
Pat Browne
Topics
 Spatial Joins
 OGC
 Themes
 Indexing
 Spatial databases
 Spatial Database Architecture
 PostGIS
Spatial Join
A spatial join associates two tables based on a
spatial relationship, rather than an the classic
non-spatial relational attribute. A spatial join
operation is used to combine two or more dataset
with respect to a spatial predicate or spatial
operation. Predicates can be a combination of
directional, distance, and topological spatial
relations (e.g. overlap, contains). In case of nonspatial join, the joining attributes must of the
same type, but for spatial join they can be of
different types.
Spatial Join Example1
Query: For all the rivers listed in the River table, find the counties
through which they pass.
SELECT r.name, c.name
FROM river AS r, country AS c
WHERE crosses(r.the_geom,c.the_geom);
The spatial predicate “Cross” is used to join River and Country tables
To view this we would add asbinary(R.the_geom,C.the_geom).
Spatial Joins
 In practice, spatial join operations are
divided into a filter step and a refinement
step to efficiently process complex spatial
data types such as point collections in a
row instance. In the filter step, the spatial
objects are represented by simpler
approximations such as their Minimum
Bounding Rectangle or Box (MBR or
MBB).
Spatial Joins Example 2
 A spatial join associate two tables based
on a spatial relationship, rather than an
attribute relationship. For example the
query:
Summarize the election results
by city.
 Could be answered using the following
SQL:
SELECT
m.name,
sum(v.fiannfail) AS FF,
sum(v.gp) AS Green,
sum(v.finegael) AS FG,
sum(v.vtotal) AS total
FROM voting_areas v, city c,
WHERE
intersects(v.the_geom, c.the_geom)
GROUP BY c.name
ORDER BY c.name;
http://www.thejournal.ie/council-elections-boundaries-2014-931336-May2013/
Dynamic and Static Data
 Static non-spatial data is usually maintained in
the table with the geometry (e.g. county name or
unique identified). In some cases this data along
with the geometry is considered immutable.
 Dynamic non-spatial data is often maintained in
a separate table e.g. land ownership.
 There can be more than one dynamic table for a
geometry table.
 Dynamic spatial can include moving objects or a
changing world (temporality requires different
treatment)
Traditional Joins
 Pure data tables can be joined with the
geometry tables for querying purposes
 A primary key is used to relate the 2 tables
together
 A primary key is a unique identifier for
each row in a table
Primary Key
ID
1
2
3
4
5
rainfall 10/7/2003 temp Hi 10/7/2003 temp Lo 10/7/2003 ID
0.5
78
68
0
80
65
1.02
76
66
0.23
81
68
0.18
80
67
1
2
3
4
5
the_geom
SRID=32140;POINT(968024.87474318 4198600.9516049)
SRID=32140;POINT(932279.183664999 4213955.37498466)
SRID=32140;POINT(952855.717021537 4223859.84524946)
SRID=32140;POINT(967568.655313907 4198112.19404211)
SRID=32140;POINT(961131.619598681 4220206.32109146)
Spatial Join1
A typical example of spatial join is “Find all pair
of rivers and cities that intersect”. The result of
join between the set of rivers {R1, R2} and cities
{C1, C2, C3, C4, C5} is { (R1, C1), (R2, C5)}.
Programming Languages
 The basic components of current languages are:
 Data types e.g. Integers, String, Polygon.
 Variables to refer to data types e.g. a = 2
 Operations on those data types e.g. area(polygon)
 Control structures e.g. sequence, iteration, and
conditions.
 Logic is an important part of programming, but it is
often implicit and external to the language. Some
languages like SQL are quite close to logic.
Data types
 A data type represents a constraint placed upon the
interpretation of data in a type system, describing
representation, interpretation, legal operations and
structure of values.
 Data types are a way to limit the kind of data that can be
used by a particular program or stored in a database
table. Types restrict the data to a certain set of values
(e.g. 1,2,3,..for Integers).
 Data types also are restricted to certain operations on
the type (e.g. addition for Integers). SQL comes with a
range of standard data types that can be used to
represent strings, integers, etc, PostgreSQL can be
extended to have additional types e.g. spatial data types.
SQL supports data abstraction
Data Abstraction- allows users to ignore
unimportant details
View Level – a way of presenting data to a
particular group of users
Logical Level – how data is interpreted when
writing queries
Physical Level – how data is manipulated at
storage level by a computer. Most users are
not interested in the physical level.
Databases use high level
declarative languages (SQL)
 Data Definition Language (DDL)
 Create, alter and delete data
 CREATE TABLE, CREATE INDEX
 Data Manipulation Language (DML)
 Retrieve and manipulate data
 SELECT, UPDATE, DELETE, INSERT
 Data Control Languages (DCL)
 Control security of data
 GRANT, CREATE USER, DROP USER
Geometry Object Model
 GOM is “abstract” (or conceptual) part of the OGC suite of
standards. It defines geometries and operations on them. The GOM
is conceptual model independent of SQL or any other language, and
hence is used to define more concrete implementations. The GOM
is based on the Object Orientation concept of inheritance, where
one class inherits attributes/methods from another. Its purpose is to
harmonise academic theory and industrial practice into a workable
consensus amongst practitioners. It has the following classes








Abstract class: Geometry
Instantiable subclasses in include:
Points which represent points in 2-dimensional space
Lines are linear edges between two points
Linestrings are connected lines (end-point is start-point of next line)
Linear Rings are ’closed’ Linestrings (last ’end-point’ is first ’start-point’)
Polygons Surface within a Linear Ring, potentially excluding inner Linear Rings
Uniform Collections of concrete Types
Spatial Types – OGC Simple
Features for SQL
 An association represents a family of links.
 Aggregation is a has-a relationship; aggregation is more
specific than association.
 Composition is a stronger variant of the "has a" association
relationship, it has a strong lifecycle dependency between
instances of the container class and instances of the contained
class(es).
 The standard does not mention UML composition, but explicitly
mentions the “owned by” black dot. Multiplicity in UML allows
to specify cardinality - i.e. number of elements - of some
collection of elements. In the standards will ill take the open
diamond to represent the part-of relation.
 Inheritance represents an is-a relation.
Spatial Types – OGC Simple
Features for SQL
Operations OGC Simple Feature Types
Operations OGC Simple Feature Types
OGC Simple Features for SQL1
 The OGC SF (similar to ISO 19125-1)
describes 2-D geometry with linear
interpolation between vertices. The simple
feature model consists of a root class
Geometry and its subclasses Point,
Curve, Surface, GeometryCollection. The
class Geometry collection has the
subclasses Multipoint, Multicurve,
MultiSurface.
OGC Simple Features for SQL1
 The OGC does not include complexes, a third
dimension, non-linear curves, `networking or
topology (i.e. connectivity information).
 Because of it relative simplicity and its support in
both the commercial & open source community
SFSQL is widely used in DBMS and is
supported in many Web applications.
 It is expected that newer more sophisticated
standards such as ISO-19107 will gradually
replace OGC SF.
OGC Simple Features for SQL (*)
 Brief description
 A simple feature is defined to have both spatial
and non-spatial attributes. Spatial attributes are
geometry valued, and simple features are based
on 2D geometry with linear interpolation
between vertices. Each feature is stored as a
row in a database table. This course covers the
OGC: GEOMETRY type with subtypes such as
POINT, LINE, POLYLINE, POLYGON, and
collections of these.
OGC Simple Features for SQL (*)
 Functionality can be described under the
1.
2.
3.
4.
following headings.
Basic Methods on Geometry
Methods for testing Spatial Relations
between geometric objects
Methods that support Spatial Analysis
Geometry Collection
OGC Simple Features for SQL (*)











Basic Methods on Geometry
Describes the dimensions and reference system (SRID) of the geometry.
Operations include Dimension, GeometryType, , conversions AsText, AsBinary, tests on geometry
include IsEmpty, IsSimple. Operations that return geometry Boundary, Envelope returns
bounding box
Methods for testing Spatial Relations between geometric objects
These polymorphic methods check relations on the generic or super class GEOMETRY and
usually return a Boolean. Main methods Equals, Disjoint, Intersects, Touches, Crosses, Within,
Contains, Overlaps, Relate( testing for intersections between the Interior, Boundary and Exterior
of the two geometries)
Methods that support Spatial Analysis
A set of geometric and ‘metric’ methods. Methods calculate distances and areas with respect to
the spatial reference system of this Geometry. Methods include Distance, Buffer, ConvexHull,
Intersection, Union, Difference, SymDifference.
Geometry Collection
A GeometryCollection is a geometry that is a collection of 1 or more geometries. All the elements
in a GeometryCollection must be in the same Spatial Reference. Subclasses of
GeometryCollection may restrict membership based on dimension and may also place other
constraints on the degree of spatial overlap between elements. Methods
NumGeometries( ):Integer—Returns the number of geometries in this GeometryCollection.
GeometryN(N:integer):Geometry—Returns the Nth
OGC Spatial Relations









Equals – same geometries
Disjoint – geometries share common point
Intersects – geometries intersect
Touches – geometries intersect at common boundary
Crosses – geometries overlap
Within– geometry within
Contains – geometry completely contains
Overlaps – geometries of same dimension overlap
Relate – intersection between interior, boundary or
exterior
OGC Spatial Operations &
Relations
Semantics of geometry
 Here are some of Euclid’s definitions from about 300 BC
1.A point is that which has no part.
2.A line is breadthless length.
3.The ends of a line are points.
4.A straight line is a line which lies evenly with the points on itself.
5.A surface is that which has length and breadth only.
6.The edges of a surface are lines.
7.A plane surface is a surface which lies evenly with the straight lines on
itself.
13.A boundary is that which is an extremity of anything
14.A figure is that which is contained by any boundary or boundaries.
 Are these definitions consistent with your understanding?
 Is a line really composed of points? How does a computer draw a line?
Semantics of geometry
 Most computer displays use raster data structures to store pixel
information.
http://en.wikipedia.org/wiki/Raster_graphics
Drawing a line in JavaScript1
Note: Code use
variable b for
y-intercept.
 The equation of a line is y = mx + c.
 dx, dy are changes in x,y
 m is the slope.
 c is y-intercept.
Drawing a line in JavaScript1
 The code on the previous slide is fine
when the slope is less than or equal to 1
(dx >= dy).
 If the slope greater than 1, the line will
have gaps (dx < dy).
 If dx = 0 then only a single point is
plotted.
 Most computer displays are pixel based.
Calculating Distance in Java
import java.util.Scanner;
public class DistanceApp {
public static void main(String[] args){
int x1, y1, x2, y2;
double distance;
Scanner scan = new Scanner (System.in);
System.out.print("Enter the x coordinate for point 1: ");
x1 = scan.nextInt();
System.out.println("The x coordinate for point 1: "+ x1);
System.out.print("Enter the y coordinate for point 1: ");
y1 = scan.nextInt();
System.out.println("The y coordinate for point 1: "+ y1);
System.out.print("Enter the x coordinate for point 2: ");
x2 = scan.nextInt();
System.out.println("The x coordinate for point 2: "+ y1);
System.out.print("Enter the y coordinate for point 2: ");
y2 = scan.nextInt();
System.out.println("The y coordinate for point 2: "+ y2);
distance = Math.sqrt( Math.pow((x2 - x1),2) + Math.pow((y2 - y1),2));
System.out.println("The distance between the two points is " + distance);
}}
Change to floating point input using nextDouble and change coord to double.
Contains Relation
Does the base geometry (small circles) contain the comparison geometry
(big circles)?
For the base geometry to contain the comparison geometry it must be a
superset of that geometry.
Geographic Information Systems and Science, ,Longley,Goodchild,Maguire,Rhind
Some spatial relations
Geographic Information Systems and Science , Longley,,Goodchild,,Maguire,Rhind
Touches Relation
Does the base geometry (small circles) touch the comparison geometry
(big circles) ?
Two geometries touch when their boundaries intersect. Raises deep
mathematical issues e.g. what is the boundary of a point?, what about
tolerance + or - a metre?
Spatial Methods






Distance – shortest distance
Buffer – geometric buffer
ConvexHull – smallest convex polygon geometry
Intersection – points common to two geometries
Union – all points in geometries
Difference – points different between two
geometries
 SymDifference – points in either, but not both of
input geometries
Convex Hull
The convex hull of a set of points is the intersection of all convex sets which
contain the points. A set of points is convex if and only if for every pair of
points p,q in S, the line segment pq is completely contained in S.
Left is convex set and right non-convex set
Convex hulls constructed around
objects.
Operations on themes
 Theme projection1: ‘selecting’ some attributes




from the countries theme. Get the Population
and Geometry of European countries
Theme selection: Name and population of
European countries with a population of 50
million or more.
Theme union : European countries with
population less than 10 million joined with those
over 10 million.
Theme overlay: See example
Theme merge : See example
Operations on themes
 Theme overlay1: Generates a new theme and
new geometry from the overlaid themes. We get
the geometric intersection of spatial objects with
the required themes. See European language
example.
 Theme merge : The merge operation performs
the geometric union of the spatial part of n
geographic objects that belong to the same
theme under a constraint condition supplied by
the user. See East/West Germany example.
Projection on Theme
Find the countries of western Europe with population greater than 50
million. This is a projection on the attribute population. Unlike a
conventional database query we often want the query result and the
original context, in this case Europe.
Theme Merge (*)
Merging two geographic objects in a selected theme (say country)
into a single object.
Theme Overlay
Anglo-Saxon
Latin languages
The lower map represents the overlay of European
countries and languages.
Indexing
 Indexing is used to speed up queries and locate
rows quickly
 Traditional RDBMS use 1-d indexing (B-tree)
 Spatial DBMS need 2-D, hierarchical indexing




Grid
Quadtree
R-tree
Others
 Multi-level queries often used for performance
(MBR)
R-tree
Example of MBR for
constructing an R–Tree
Index of polygons
Minimum Bounding Rectangles
Minimum
Bounding
Rectangle
Study
Area
Spatial enabled DB Overivew
 Database – an integrated set of data on a
particular subject. Can include spatial and nonspatial and possible temporal.
 Databases offer many advantages over files
 Relational databases dominate for non-spatial
use, object-relational databases (ORDBMS)
often used for spatial data.
 Databases address some limitations for GIS
Guting’s1 definition of a spatial
database
 (1) A spatial database system is a
database system
 (2) It offers spatial data types in its data
model and query language
 (3) It supports spatial data types in its
implementation, providing at least spatial
indexing and efficient algorithms for
spatial join2.
Why use a database for GIS?
 A GIS is strictly speaking not a database




system. A GIS can be connected to a DBMS.
A GIS cannot efficiently manage large quantities
of non-spatial data (e.g. at government
department level).
They lack ad hoc querying capability (they
provide a restricted form of predefined queries)
They lack indexing structures for fast external
data access (they use in memory techniques).
They lack a 'logic' (e.g. first order logic of the
relational calculus)
Why use a database for GIS?
 Databases offer the following functions:
 Reliability
 Integrity: enforces
consistency
 Security
 User views
 User interface
 Querying
 Updating
 DB theory has a
Mathematical basis







Data independence
Data Abstraction
Self-describing
Concurrency
Distributed capabilities
High performance
Supports spatial data
types using ADTs.
 Alternative: files
Choice database for GIS?
 Choice of DBMS:
 Commercial (Oracle, DB2) or
 Open source (PostgreSQL, MySQL) .
 We will PostgreSQL with PostGIS spatial
extensions.
 PostgreSQL is an Object Relational
Database System (ORDBMS).
Database Architecture for GIS(*)
 Pure Relational Approach
 Spatial data can be stored in a pure RDBMS. The coordinates for the spatial
data can be stored in tables. Uses existing technologies, requires no
additional software (for the pure DBMS perspective).
 Drawbacks
 It is difficult to represent and query complex spatial structures (such as a polygon
with holes) or topological relationships ( network connectivity, polygon
adjacency). No ordered lists.
 Violates independence principle, user must know about data storage. Change of
geometric representation requires deep reorganization of the database and
query formulation.
 Poor performance, requires a lot of processing of the relational tuples that
represent the spatial information.
 Lack of user friendliness because users have to manipulate tables of points.
 Difficulty of defining new spatial types.
 The impossibility of expressing geometric computations such as adjacency tests,
point query, or window query.
Database Architecture for GIS(*)
 Loosely Coupled
 Many commercial and government agencies use this





approach.
Uses a RDBMS to store 'attribute' or descriptive
information e.g. the name of a road not its geometry.
A specific module for spatial data management.
Drawbacks
The coexistence of heterogeneous data models, which
implies difficulties in modeling use and integration.
A partial loss of DBMS techniques e.g. recovery,
querying, optimization.
Database Architecture for GIS(*)
 Loosely Coupled Architecture
Application Programs
Relational DBMS
Geometric Processing
(GIS)
Database
Files
Database Architecture for GIS(*)
 Integrated
 DBMS extensibility address many of the problems inherent in
the RM and LC approaches.
 Most commercial databases that offer facilities to handle spatial
data (PostgreSQL, Oracle, DB2) take this approach. The basic
idea is to add new types and operations to the RM as follows:
 The query language is extended to manipulate spatial data as
well as descriptive data. New spatial types (point, line, and
polygon) are handled as basic types by the DBMS.
 Many other DBMS functions such as query optimization, are
adapted in order to handle geo-spatial data efficiently.
 Drawback: Does not provide full GIS functionality (cartography).
We must use additional software such as Geoserver to make
(or render) an attractive map from the raw vectors stored in the
DBMS.
What can PostGIS do?
Many PostGIS functions available via SQL
Compliant with OGC1 Simple Features Specification
Coordinate
transformation
Crosses
Identify (SRID)
Contains
Buffer
Area
Touches
Length
Crosses
Point on surface
Within
Return geometry
as SVG
Overlaps
Contains
Overlaps
What can PostGIS do?
PostGIS supports a geometry type which is
compliant with the OGC standard for Simple
Features.






POINT( 50 100 )
LINESTRING ( 10 10, 20 20 )
POLYGON ( ( 0 0, 5 5, 5 0, 0 0 ) )
MULTIPOINT ( ( 1 1 ), ( 0 0 ) )
MULTILINESTRING ( … )
MULTIPOLYGON ( … )
HOW Spatial Databases Fit into
GIS
LAN
Internet
Editing
Loading
Analysis
Mapping
Database
Features
GIS
Image from Paul Ramsey Refractions Research
GIS
Web
Client
Other
ProstgreSQL
 PostgreSQL itself provides the main
features of a RDBMS. Includes other
advanced features such as:






Inheritance
Functions
Constraints
Triggers
Rules
Transactional integrity
 Permits an ‘OO like’ style of programming
PostgreSQL/PostGIS
The data is stored in a relatively simple format with the
attributes and geometry stored in a single table. It can be
viewed as Well Known Text (WKT) or displayed
graphically using the asBinary(the_geom) function.
Spatial
reference Data type Coordinates
number
Attribute Data
name
Brio Refining
Crystal Chemical
North Cavalcade
Dixie Oil Processors
Federated Metals
city
Friendswood
Houston
Houston
Friendswood
Houston
hrs
50.38
60.9
37.08
34.21
21.28
status
active
active
active
active
active
st_fed
the_geom
Fed
SRID=32140;POINT(968024.87474318
4198600.9516049)
Fed
SRID=32140;POINT(932279.183664999
4213955.37498466)
Fed
SRID=32140;POINT(952855.717021537
4223859.84524946)
Fed
SRID=32140;POINT(967568.655313907
4198112.19404211)
State
SRID=32140;POINT(961131.619598681
4220206.32109146)
How does it work?
 Spatial data is stored using the coordinate
system of a particular projection.
 That projection is referenced with a Spatial
Reference Identification Number (SRID)
 This number relates to another table
(spatial_ref_sys) which holds all of the
spatial reference systems available.
 This allows the database to know what
projection each table is in, and if need be, reproject from those tables for calculations or
joining with other tables.
Coordinate Projection
SRID=3005;MULTILINESTRING((1004687.04355194594291.
053764096,1004729.74799931 594258.821943696))
NORTH AMERICAN DATUM OF 1983
SRID 3005=NAD83
SRID 4326=WGS84
SRID=4326;MULTILINESTRING((125.934150.364070000000
1,-125.9335 50.36378))
Coordinates of one table can be converted to those of another table. This
permits the ‘geometry’ in each table to match. Relatively easy to do in PostGIS
Spatial Database Components
 The Geometry metadata table
table
schema
geometry
column
coord
dim
table name
brazos
texas_counties
the_geom
2
32139
MULTIPOLYGON
brazos
texas_rivers
the_geom
2
32139
MULTILINESTRING
brazos
texas_roads
the_geom
2
32139
MULTILINESTRING
brazos
tx_maj_aquifers
the_geom
2
32139
MULTIPOLYGON
brazos
tx_min_aquifers
the_geom
2
32139
MULTIPOLYGON
brazos
txzip_codes
the_geom
2
32139
MULTIPOLYGON
brazos
bz_landmarks
the_geom
2
32139
POINT
srid
type
spatial_ref_sys
 postgis=# \d spatial_ref_sys
Table "public.spatial_ref_sys"
Column
|
Type
| Modifiers
-----------+-------------------------+----------srid
| integer
| not null
auth_name | character varying(256) |
auth_srid | integer
|
srtext
| character varying(2048) |
proj4text | character varying(2048) |
Indexes:
"spatial_ref_sys_pkey" PRIMARY KEY, btree (srid)
geometry_columns
 postgis=# \d geometry_columns
Table "public.geometry_columns"
Column
|
Type
| Modifiers
-------------------+------------------------+----------f_table_catalog
| character varying(256) | not null
f_table_schema
| character varying(256) | not null
f_table_name
| character varying(256) | not null
f_geometry_column | character varying(256) | not null
coord_dimension
| integer
| not null
srid
| integer
| not null
type
| character varying(30) | not null
Indexes:
"geometry_columns_pk" PRIMARY KEY, btree
(f_table_catalog, f_table_schema, f
_table_name, f_geometry_column)
Database Rules
 Rules help prevent human error when
modifying a data set
 Rules are user defined
 Rules are such things as;
 “A fire hydrant must be located on a water
line”
 Rivers should flow down hill.
Constraints
 Constraints are similar to rules, but are
less assertive.
 Constraints are provided by the DBMS
and are applied by the user
 A Constraint would be “Parcel_ID Not Null”
- meaning a number ID has to be provided
when a parcel is created.
Constraints
Constraint
GIS examples
Two spatial objects cannot exist at the same point
Uniqueness
Non-Null
All Address points must have co-ordinates
Range
All heights in Ireland must be in range -100 to 2000 metres
Relationship
Every river must be connected to the sea, a lake or other
river (Can rivers cross)
Cardinality
Each side of a triangle has a 1:2 relation with the others
Inclusion
All counties are polygons
Covering
A boundary may be a townland and/or a barony.
Disjointedness
All roads must be only a primary or a secondary or a
regional
Referential Integrity
A county border must be represented by a ground feature
Geometrical
Triangles must have three sides
Orientation
Roads are usually to the front of houses
Topological
Inner walls must be "inside" buildings
General
Complex rules built from above constraints
Constraints
How can we define in front of?
Data integrity
Valid
Invalid
select count(*) from voting_areas
where not isvalid(the_geom);
Temporal Example: roads, buildings,
and regions
Consider a line. From the properties of metric spaces it has a length.
Temporal Example: roads, buildings,
and regions
A
B
Lets call it a road. From graph theory we have a path
Temporal Example: roads, buildings,
and regions
A
F1
B
Purple line segment
represents both a road
and a fence.
Lets add a field (F1) with an area and a topology.
Example: roads, buildings, and regions
A
F1
B
Lets add an administrative region (outer red rectangle) and some houses
Example: roads, buildings, and regions
A
F2
F3
B
Lets divide the field in two by inserting a new fence.
We need to delete the old area and add two new areas.
What about adjacency relation between fields?
Example: roads, buildings, and regions
Time1
Time2
A
A
F1
F2 F3
B
B
Imagine a picture of the world at Time1 and Time2.
Not only have some objects changed but some spatial
relationships have changed.
An addition can induce a deletion and a deletion can induce
an insertion.
Example of temporal queries
• Is there a route from A to B? (now is
•
•
•
•
assumed)
Was there a route from A to B in Time1?
Does the route in query 1 pass through
the administrative region?
Does the route in query 1 pass touch the
administrative region ?
What fields were adjacent to F2 in
Time2?
Raster Image Data Not Covered