A:ribute Tables and Tabular Data. James Frew ESM 263 Winter

Transcription

1 A:ribute Tables and Tabular Data 1

2 Outline A:ribute tables Joins and relates RelaEonal databases ImporEng tabular data 2

3 A:ributes stored/organized in tables row = spaeal feature column = a:ribute row i column j = value of a:ribute j for feature i 3

4 Rendering the A:ribute Table in spreadsheet in ArcMap 4

5 Table CharacterisEcs All tables Row order doesn t ma:er rows can be ordered on any column value(s) Columns are typed ArcGIS: integer, real, text, date Feature a:ribute tables 1 row per feature table row feature ID geometry object 1 table per feature class shapefile, coverage, geodatabase feature class,... 5

6 A:ribute ManipulaEon in ArcGIS

7 A:ribute Querying in ArcGIS: Select by A:ributes simplified SQL query columns: all table: current constraint: column {op} value choose from available or specify directly not quite a database! 7

8 ConnecEng Tables What if a:ributes are in more than 1 table? Create rela.onships between tables cardinality: #rows (table1) #rows (table2) 8

9 ConnecEng Tables in ArcGIS Connect tables using common key values join: concatenates relate: links (but keeps separate) ArcGIS, not database terminolgy! 9

10 Joins 1 to 1 Many to 1 10

11 Relate 1 to many e.g. one soil map unit two soil components component NB: can t join 1-to-many: why? would have to replicate features 11

12 Summary: Tables in ArcGIS ArcGIS isn t quite a database Enforces 1 feature 1 a:ribute table row Joins and relates are part of map, not data Can t query muleple tables simultaneously Have to explicitly join or relate them first Can t save queries and run them later 12

13 What s a Database? A colleceon of data? e.g. all the data for your group project A computer informaeon system? file? spreadsheet? GIS Date s criteria: a database enables/enforces: integraeon sharing persistence eneees, a:ributes, and relaeonships» Date, C.J., An introduceon to database systems (7th ed.) ISBN

14 Tables in a Database Table = en.ty e.g. professors Row = instance of an enety e.g. Frew also called: tuple Column = a7ribute of an enety e.g. shoe size A database is really picky about what you put in a table... 14

15 RelaEon (Table) Rules Only one value in each cell (interseceon of row and column) All values in a column are about the same subject Each row is unique Column order doesn t ma:er Row order doesn t ma:er 15

16 Usually Need More than 1 Table Avoid redundancy: if single table, then a:ribute values shared by >1 instance must be repeated in each instance e.g. 58 students taking meets in Bren student records have Bren 1414 as meeeng place what happens when class moves? Consequences of redundancy more sensieve to typos and transcripeon errors fragile updates: have to change muleple copies 16

17 How Databases Use MulEple Tables Eliminate redundancy by normalizing single table into muleple tables Each table = single kind of thing Each row = single thing Preserve relaeonships by references between tables Collapse redundant a:ributes into single key (a:ribute shared between tables) RelaEonships implied by matching key values 17

18 Keys A key uniquely idenefies, and can therefore be used as a reference to, a single row Primary key a:ribute whose value uniquely idenefies a row Data values that are naturally unique may be more than 1 a:ribute Arbitrary/syntheEc value e.g. auto-incremeneng counter Foreign key a:ribute whose value corresponds to another row s (usually in another table) primary key Foreign keys are how databases maintain explicit rela.onships between rows, within or between tables 18

19 NormalizaEon Step 0: non-normalized single table PIN Owner Owner Address Sale date Acres Zone code Zoning P101 Wang Chang 101 Oak St. 200 Maple St residential P102 Smith Jones 300 Spruce St. 105 Ash St commercial P103 Costello 206 Elm St commercial P104 Smith 300 Spruce St residential 19

20 NormalizaEon First Normal Form no mule-valued a:ributes PIN Owner Owner address Sale date Acres Zone code Zoning P101 Wang 101 Oak St residential P101 Chang 200 Maple St residential P102 Smith 300 Spruce Rd commercial P102 Jones 105 Ash St commercial P103 Costello 206 Elm St commercial P104 Smith 300 Spruce Rd residential 20

21 NormalizaEon Second normal form separate eneees: parcel, owner, address 21

22 NormalizaEon Third normal form separate remaining non-key dependencies 22

23 Advantages of Databases over Files Avoids redundancy and duplicaeon Reduces data maintenance costs ApplicaEons are separated from the data ApplicaEons persist over Eme Support muleple concurrent applicaeons Be:er data sharing Security and standards can be defined and enforced 23

24 PotenEal Disadvantages of Databases over Files Expense but: good free ones: SQLite, PostgreSQL Performance especially complex data types but: databases are really good at using cached and/or sorted copies of their data to speed up queries space-for-speed tradeoff IntegraEon with other systems can be difficult not so much of an issue any more: most systems that could use databases know how to connect to them 24

25 References Chang, K.T., IntroducEon to Geographic InformaEon Systems, 5 th ed. ISBN X Date, C.J., An IntroducEon to Database Systems, 7 th ed. ISBN