Geography Data Type
New to SQL Server 2008 are the spatial data types called Geography and Geometry. The Geography data type can store information for areas and points on the earth. It also provides a built-in function to calculate distance and overlaps with other locations. This data type stores and handles calculations based on round-earth (or ellipsoidal) data, which relates to coordinate systems such as GPS and longitude-latitude.
Databases have been storing positional data for years, like the sample data below where we see two dedicated fields to store longitude and latitude for each JProCo location. Please note that your if you ran the SQLArchChapter5.0Setup.sql script then the Location table in the figure below does not currently have the latitude and longitude fields. This is our first goal we will do very soon after some more explanation.
If you were to ask this table in a query what the longitude-latitude difference between Seattle and Boston is, it would have no idea. To the query, these are just arbitrary flat data values in a table. In order to turn each pair of numbers into a meaningful geographical point, we would have to extract this data into a custom application outside of SQL Server (e.g., into a C# application) and then run calculations using the customized app.
In other words, the true ellipsoidal nature of this geographical data wasn’t stored in the database prior to SQL Server 2008. The database couldn’t have differentiated our sample longitude-latitude data above from any other kind of information in the database. Now, thanks to SQL Server 2008, you no longer need a separate custom application, because both of these pieces of data can be stored in one Geography field. As well, the Geography type makes available all the built-in functionality, to perform calculations involving round-earth data, which would have been contained in a custom app.
Storing Latitude and Longitude
Prior to SQL Server 2008, two float or decimal fields would be used to house latitude and longitude. Now you can store these, as well as other geospatial data, in one Geography field.
Let’s begin by looking at all the records in our Location table, as well as its design. We see typical location data – city, state, street – and the five JProCo office locations (Seattle, Spokane, Chicago, Boston, and Philadelphia). Into this table we will add fields for latitude and longitude, and then we’ll see how to combine those into one Geography field..
Let’s start by adding latitude and longitude fields and populate these fields using the values based on the code below:
ALTER TABLE Location
ADD Latitude FLOAT NULL
ALTER TABLE Location
ADD Longitude FLOAT NULL
SET Latitude = 47.455, Longitude = -122.231
WHERE LocationID = 1;
SET Latitude = 42.372, Longitude = -71.0298
WHERE LocationID = 2;
SET Latitude = 41.953, Longitude = -87.643
WHERE LocationID = 3;
SET Latitude = 47.668, Longitude = -117.529
WHERE LocationID = 4;
SET Latitude = 39.888, Longitude = -75.251
WHERE LocationID = 5;
Creating Geography as a Field in a Table
We’re now going to add another field called GeoLoc (short for geographical location), which will use the new Geography data type. We have the latitude and longitude fields populated with values for each JProCo location. The Geography field GeoLoc has also been added to the table but is not yet populated.
Populating a Geography Data Type
Based on the two data points, latitude and longitude, we can generate the geospatial locations for the GeoLoc field. Use the Point-static function called GEOGRAPHY to pass in the latitude and longitude values along with a style specifier value (4326 is the standard which is used the most).
SET GetLoc = GEOGRAPHY::Point(Latitude , Longitude , 4326)
The style specifier we used to format our Geography value is also known as a spatial reference identifier or SRID and identifies which spatial reference system the coordinates belong to. The SRID 4326 represents WGS 84, which is the most commonly used system and is used by many GPS systems.
Calculating Distance between two points on the earth
We have successfully combined latitude and longitude into the GeoLoc field. However, since the values in the GeoLoc column are a little cryptic to read by the human, we can create a variable and capture each city’s GeoLoc value into its respective variable.
STAsText( ) is one of the Spatial Type methods you can use with the Geography type. The STDistance( ) method calculates the shortest distance (in meters) between two Geography data points. To have STDistance( ) return the distance from Seattle to Boston in kilometers (KM), we have divided the returned value by 1000). Without this step, the result is just over 4 million meters (4,014,163 meters). Calculating the distance from Seattle to Boston, which is just over 4,014 KM.
Note: If you want to setup the sample JProCo database on your system you can watch this video. For this post you will want to run the SQLArchChapter5.0Setup.sql script from Joes 2 Pros Volume 3.
The STDistance function of the Geography Data type calculates the distance between two points in …
Do not forget to participate in special question over here: Pluralsight Giving Away Free Subscription to Quiz Participants
Please leave your answer in comment section below with correct option, explanation and your country of resident.
Every day one winner will be announced from United States.
Every day one winner will be announced from India.
A valid answer must contain country of residence of answerer.
Please check my facebook page for winners name and correct answer.
Winner from United States will get Joes 2 Pros Volume 3.
Winner from India will get Joes 2 Pros Volume 3.
The contest is open till next blog post shows up at http://blog.sqlauthority.com which is next day GTM+2.5.
Reference: Pinal Dave (http://blog.SQLAuthority.com)