Transcript Attribute Data - Computer Science

Attribute Data
CampusID
Name
Type
6
Murphy
Academic
2
2001
9
Hopkins
Support
2
946
12
Maintenance
Support
1
1848
15
Hickey
Support
2
2367
17
Shay-Loughlen
Dorm
3
1298
CS 128/ES 228 - Lecture 6a
Floors
Footprint
1
Why Databases?

One of the advantages of vectorbased data is that each datum
corresponds to a (portion) of a “real”
object.

BUT… Objects are more than
geographic locations
CS 128/ES 228 - Lecture 6a
2
Database “Review”
 Databases consist of tables
 Each table holds records
 Records are in rows
 Each record consists of fields, i.e.
individual data items
 Fields are in columns
 Fields that have unique values within a
table are called “keys”
CS 128/ES 228 - Lecture 6a
3
Sequential Databases
Once upon a time,
computer scientists tried to
put all the data for an
application in one table
It didn’t work
CS 128/ES 228 - Lecture 6a
4
Flaws of Sequential Databases

Not all data is homogeneous

Monolithic structure leads to
contention problems

Separation of data into “sub”
databases leads to duplication
(and inconsistencies)
CS 128/ES 228 - Lecture 6a
5
Relational databases
 Use many (related) tables
of data, with minimal
duplication
 Tables are “linked” through
common values in particular
fields
 “Queries” permit rich “data
mining”
CS 128/ES 228 - Lecture 6a
6
Relational Databases in GIS
One table traditionally holds
geographic information
 Other tables hold data about other
attributes
 Tables are linked through “object
Ids”
 (Object Ids should be independent
of the software)

CS 128/ES 228 - Lecture 6a
7
Relational Databases in GIS

Consider a “Buildings” layer


“GIS” data includes location (including
coordinate system), symbology,
internal bookkeeping
“External” data includes name of
building, date built, purpose, etc.
CS 128/ES 228 - Lecture 6a
8
Sometimes they are mixed
CS 128/ES 228 - Lecture 6a
9
Queries

Information is
gained by
linking tables
through “joins”

Queries can
involve
computed
quantities, etc.
CS 128/ES 228 - Lecture 6a
10
Time out for a query demo
CS 128/ES 228 - Lecture 6a
11
Storing Topology

Shapefiles contain no topological
information

But topology is important
SOLUTION

Store the topology elsewhere
CS 128/ES 228 - Lecture 6a
12
Data Structure for Topology
Zone Zone Bou
1
2
nd.
2
1
B
4
A
10
C
8
D
7
B
4
A
C
5
B
C
10
C
D
8
3
5
9
A
6
CS 128/ES 228 - Lecture 6a
13
Practicality
 GIS must maintain topology (to be
useful)
 GIS can’t store topology (in
shapefiles)
 SO, GIS must compute topology
CS 128/ES 228 - Lecture 6a
14
Topological Operations

Merge two regions into one
A
C
B
CS 128/ES 228 - Lecture 6a
15
Topological Operations

Split one region into two
B
A
C
CS 128/ES 228 - Lecture 6a
16
Topological Operations

Clip an image
CS 128/ES 228 - Lecture 6a
17
Topological Operations

Erase an image
CS 128/ES 228 - Lecture 6a
18
Other operations

Union

Intersect

“Snaps”
CS 128/ES 228 - Lecture 6a
19
How this is used

Find nearest
neighbor

Find largest
“empty” region
CS 128/ES 228 - Lecture 6a
20
What distinguishes (some) GIS’s

The types of these operations that
they can perform.
Both Database and Topological
CS 128/ES 228 - Lecture 6a
21