Geodatabase 101 Why, What, & How Beau Dealy Dealy Geomatics, LC beau@dealygeo.com Curt Moore InfiniTec, Inc. cmoore@infinitec.net
... first, a brief explanation. Geodata traditionally stored as two components Spatial data usually specialized/proprietary Attribute data tabular file or RDBMS (that's a Relational Database Management System, like SQL Server, Oracle, DB2, or PostgreSQL) Examples ESRI's Coverage & shapefile formats. A Geodatabase (GDB) is A database with extensions for storing, querying, and manipulating geographic information and spatial data. (Wikipedia, 2007)
Why? Why should you be considering GDB? Centralization Scalability Security Rules c
Centralization All data is stored in an RDBMS No more herding cats to get the right files to the right folks. Easier to back-up and make redundant copies Example 3 data formats viewed in ArcCatalog.
Centralization (cont'd) Same three formats viewed in OS file manager (Windows Explorer) Coverage Shape Geodatabase
Scalability In ESRI-land, one editor and a 2GB limit and for personal GDB For RDBMS: multi-user edits and your RDBMS is only limiting factor (usually advertised as 'Unlimited') Unlimited = storing raster data in GDB is possible (and a good thing).
Security Access Restrictions Only the folks that should be looking at or editing data are allowed to do so. Management Using RDBMS gives the opportunity for your IT department/vendor's DBA to do his job. (protect, manage, backup)
Rules Establishing rules prevents inconsistencies and errors in spatial and tabular data. Puts tabular data entry 'on rails' Topology Rules of relationship between spatial objects. Example: Moving a parcel line means lot, block, and possibly subdivision boundaries must move as well.
What? What What are the components of a GDB? can be done with a GDB that I can't already do with my current geodata structure? What is this going to cost me?
What are the components of a GDB? DISCLAIMER - My experience is with ESRI (ymmv) Components of an ESRI GDB (ESRI 2007) Tables (non-spatial (attribute) data) Feature Classes (spatial data) Feature Datasets (feature class containers) Relationship Classes (link tables) Topologies (spatial validation rules)
What can be done that I can't already do? Relationships persist in the GDB Link/join not limited to desktop software Relates = better performance than one massive table Topology Shape files are dumb (think: editing adjacent polys in a shapefile) Establishing topological rules saves time in the long-run Integration Putting geodata in an RDBMS opens opportunities to integrate with existing database applications (e.g. CAMA, tax, deeds, etc.)
What is this going to cost me? 'Going geodatabase' doesn't have to be expensive. Existing software may be enough to start ArcGIS 9.2 'File Geodatabase' uses MS SQL Express (bundled) You may already be paying for part of the solution Have a RDBMS? Only need to budget extra money for the middleware (e.g. ARCSDE) Open-source (free) solutions are available to those with time & know-how (e.g. PostGIS)
How? How to start migrating data? How to get up-to-speed on best practices? How do I use it?
How to start migrating data? Start with data that changes little Clean up your mess! Using an RDBMS? Import aerial photography Layers from outside organizations (no editing) Define all projections, remove unused fields, create metadata Now is the time to make changes to your data model If you've got folks that depend on your current database design, you may have to run parallel sets for a while anyway.
How to get up-to-speed on best practices? Talk to peers Conferences User Groups Message boards A vendor can be a peer as well Do your homework White papers Web Seminars Read case studies Download & study data and data models
How do I use it? Demo (ArcGIS Desktop, ArcGIS Server). Adding data Stand-alone feature data vs. feature data sets Using ArcSDE: Desktop -> Server -> Web Questions?