Tutorials

The following tutorials provide step by step instructions in C++, Java, Javascript and Python languages to get started writing applications using GPUdb. GPUdb provides a rich set of APIs for querying data. These guides demonstrate only a small set of the available API. A detailed description of the complete interfaces is available under API Reference.

• C++ Tutorial
  • Downloading GPUdb API
  • Connecting to GPUdb
  • Loading Data
  • Retrieving Data
  • Running Queries
  • Complete Sample
• Java Tutorial
  • Downloading GPUdb API
  • Connecting to GPUdb
  • Loading Data
  • Retrieving Data
  • Running Queries
  • Complete Sample
• JavaScript Tutorial
  • Downloading GPUdb API
  • Connecting to GPUdb
  • Loading Data
  • Retrieving Data
  • Running Queries
  • Complete Sample
• Node.js Tutorial
  • Downloading the GPUdb API
  • Connecting to GPUdb
  • Loading Data
  • Retrieving Data
  • Running Queries
  • Complete Sample
• Python Tutorial
  • Downloading GPUdb API
  • Connecting to GPUdb
  • Loading Data
  • Retrieving Data
  • Running Queries
  • Complete Sample

Exporting Data

The data in GPUdb tables can be exported by the GPUdb Admin web interface. Currently only csv format is supported. To export the data, navigate to the GPUdb Admin URL http://<GPUdb-head-node-ipaddress>:8080/gadmin . On the main toolbar at the very top, click on Data to bring up a list of all the tables. Click on the name of the table to be exported, which will display the table data. On the top of the data table, click on the Export button to bring up export data dialog. Make the appropriate selection based on the desired rows to be exported and click Export. The browser should now display a system dialog prompting for the location and name of the file in which to save the exported data.

Geospatial Capabilities

GPUdb has support for geospatial processing and visualization. Currently supported geospatial objects include geometric objects and tracks.

Geometric Objects

Geometric objects include points, polygons, linestrings and multi-versions of each. To create a geometric object, include a string column named 'WKT' in your schema definition. Now you can load data in WKT (Well-Known Text) format. Internally GPUdb will convert the WKT string representation into a geometric object.

Example WKT Strings:

• POINT (5 10)
• LINESTRING (1 2, 3 4)
• POLYGON ((5 10),(10 10),(10 5),(5 5),(5 10))

Tracks

Tracks in GPUdb consist of geospatial data (i.e. latitude and longitude) along with grouping data (i.e a track id) and ordering data (i.e. timestamp).

To support track processing, in your schema definition include the following columns:

• a numeric 'x' column representing longitude
• a numeric 'y' column representing latitude
• a numeric 'TIMESTAMP' column representing time (the ordering column)
• a string 'TRACKID' column (the grouping column)

Geospatially-Aware Endpoints

The following endpoints are optimized for geometric objects and tracks:

• /aggregate/convexhull
• /filter/byarea
• /filter/bybox
• /filter/bygeometry
• /filter/byradius
• /filter/byseries
• /get/records/byseries
• /visualize/image
• /visualize/image/classbreak
• /visualize/video