Introduction to Spatial SQL

Introduction to Spatial SQL

Matt FancherGIS Specialist

Public Utilities Commission of Ohio

2012 Ohio GIS ConferenceSeptember 19 - 21, 2012 | Hyatt Regency Hotel | Columbus, Ohio

Structured Query Language

SQL can select data from a database

SQL can insert data in a database

SQL can update data in a database

SQL can delete data from a database

“I want to SELECT some informationFROM a particular source

WHERE certain criteria are met.”

SELECT column1, column2FROM table1

WHERE criteria are met

“I want to SELECT some informationFROM a combination of sourcesWHERE certain criteria are met.”

SELECT table1.column, table2.columnFROM table1 JOIN table2ON table1.id = table2.idWHERE criteria are met

DECLARE @variable AS data type

SET @variable = some value

SET @variable = some object

SELECT column1, column2FROM table1

WHERE column1 = @variable

SpatialSQL

GeographyGeometry

Integer Text

VarcharDate

Time

Blob

Decimal

Long

Char

Double

Short

Numeric

Float Bit

Money

Binary

Datetime

Real

Well Known Text

Point(x y)

Line(x1 y1, x2 y2, … , xn yn)

Polygon(x1 y1, x2 y2, … , xn yn, x1 y1)

STPointFromText(‘Point(x y)’, SRID)

STLineFromText(‘Line(x1 y1, x2 y2, … , xn yn)’, SRID)

STPolygonFromText(‘Polygon(x1 y1, x2 y2, … , xn yn, x1 y1)’, SRID)

Geography::STPointFromText(‘Point(x y)’, SRID)

Geometry::STPointFromText(‘Point(x y)’, SRID)

“ I want to INSERT a new record INTO a database table

using this list of VALUES.”

INSERT INTO LocationTable VALUES(101, geography::STPointFromText(‘Point(x y)’, SRID))

Expose Spatial PropertiesTest Spatial Relationships

Perform Spatial Operations

STX & STYSTLength()

STArea()STCentroid()STEnvelope()

Spatial Properties

SELECT LakeNameFROM Lakes

WHERE Shape.STArea > 1000000

STIntersects()STDisjoint()STDistance()STTouches()STWithin()

Spatial Relationships

SELECT Congress.District, County.CountyFROM CongressJOIN County ON

Congress.Shape.STIntersects(County.Shape) = 1

STIntersection()STUnion()

STDifference()STBuffer()

STConvexHull()

Spatial Operations

DECLARE @Township as Geometry;DECLARE @City as Geometry;DECLARE @Difference as Decimal;

SELECT @Township = Shape FROM TownshipWHERE Name = ‘Orange Township’;

SELECT @City = Shape FROM CityWHERE Name = ‘Columbus’;

SELECT @Difference = @Township.STDifference(@City).STArea;

Address Google MapsGeocode Service

Latitude/LongitudeCoordinates

SQLStatements Result Set Display on Page

Application Process:

SELECT Electric_Company, Company_Type, ChoiceFROM Electric_Company_TableWHERE Shape.STIntersects(geography::STPointFromText('POINT(<longitude> <latitude>)', 4269)) = 1;

SELECT Municipality_NameFROM Municipal_Utility_TableWHERE Shape.STIntersects(geography::STPointFromText('POINT(<longitude> <latitude>)', 4269)) = 1;

SELECT Telephone_Company, Telephone_ExchangeFROM Telephone_Exchange_TableWHERE Shape.STIntersects(geography::STPointFromText('POINT(<longitude> <latitude>)', 4269)) = 1;

SELECT County_NameFROM County_TableWHERE Shape.STIntersects(geography::STPointFromText('POINT(<longitude> <latitude>)', 4269)) = 1;

--First create a point from the user’s input

DECLARE @Point AS geometry;SET @Point = geometry::STPointFromText('POINT(<x_coord> <y_coord>)', 0);

--Then buffer the point by a specified distance

DECLARE @Buffer AS geometry;SET @Buffer = @Point.STBuffer(<linear_distance>);

--Finally execute a select statement to estimate the population in the buffer

SELECTROUND(SUM(PopDen * Shape.STIntersection(@Buffer).STArea() / 27878400),0)AS PopulationEstimateFROM Census_Block_2010WHERE Shape.STIntersects(@Buffer) = 1;

SQL Solution:

The End

Web Server

ColdFusion Script

T-SQLStatements

Web MapBuilt on

Google MapsAPI

MS SQL ServerDatabase

Request

Request Result Set

Response

T-SQL Script Implementation:

Loading Spatial Data

• Shape2sql:

SELECT

P_Id LastName FirstName Address City1 Hansen Ola Timoteivn 10 Sandnes2 Svendson Tove Borgvn 23 Sandnes3 Pettersen Kari Storgt 20 Stavanger

SELECT LastName, FirstName FROM Persons

The "Persons" table:

LastName FirstNameHansen OlaSvendson TovePettersen Kari

The result-set will look like this:

WHERE

P_Id LastName FirstName Address City1 Hansen Ola Timoteivn 10 Sandnes2 Svendson Tove Borgvn 23 Sandnes3 Pettersen Kari Storgt 20 Stavanger

The "Persons" table:

SELECT * FROM Persons WHERE City='Sandnes'

P_Id LastName FirstName Address City1 Hansen Ola Timoteivn 10 Sandnes2 Svendson Tove Borgvn 23 Sandnes

The result-set will look like this:

JOINP_Id LastName FirstName Address City1 Hansen Ola Timoteivn 10 Sandnes2 Svendson Tove Borgvn 23 Sandnes3 Pettersen Kari Storgt 20 Stavanger

O_Id OrderNo P_Id1 77895 32 44678 33 22456 14 24562 15 34764 15

The "Orders" table:

The "Persons" table:

LastName FirstName OrderNoHansen Ola 22456Hansen Ola 24562Pettersen Kari 77895Pettersen Kari 44678

The result-set will look like this:

SELECT Persons.LastName, Persons.FirstName, Orders.OrderNoFROM PersonsINNER JOIN OrdersON Persons.P_Id = Orders.P_IdORDER BY Persons.LastName

Overview

• Agenda:– Native spatial data types in RDBMS– Structure Query Language (SQL) syntax– Examples of spatial SQL use

• Notes:– Microsoft SQL Server 2008 used in examples– Given from perspective of ArcGIS Specialist

Native Spatial Data Types in RDBMS

• Analogous to integer, text, date, etc.• Used to store spatial data objects: points, lines,

and polygons• Intrinsic to the RDBMS; no extension needed• MS SQL Server has to spatial data types:

– Geography (i.e. for geographic coordinates)– Geometry (i.e. for projected coordinates)

• Spatial data implementation compliant with OGC Simple Feature Access standard

Introduction to Spatial SQL

A brief primer on working with the native spatial data types in Microsoft

SQL Server