Hortonworks Docs » DataFlow 3.5.1 » Building a SAM Application

Building a SAM Application

Also available as:

PDF

loading table of contents...

Building an Application

Building an Application

Prerequisites

You have integrated SAM
You have set up appropriate environments and service pools

Launch the Stream Builder UI
Add a New Stream Application
Add a Source
As described in the HDF Overview, Stream Builder offers four types of builder components: sources, processors, sinks, and custom components. Start building your application by adding a source.
Connect Components
Once you have added and configured your source, add additional processors and sinks to the canvas. To pass a stream of events from one component to the next, create a connection between the two components. In addition to defining data flow, connections allow you to pass a schema from one component to another.
Join Multiple Streams
Joining multiple streams is an important SAM capability. You accomplish this by adding the Join processor to your stream application.
Filter Events in a Stream
You can use SAM to filter events in the stream. You accomplish this by using Rule processor, which translates rules into SQL queries that operate on the stream of data.
Use Aggregate Functions over Windows
Windowing is the ability to split an unbounded stream of data into finite sets based on specified criteria such as time or count, so that you can perform aggregate functions (such as sum or average) on the bounded set of events. In SAM, you accomplish this using the Aggregate processor. The Aggregate processor supports two window types, tumbling and sliding windows. You can create a window based on time or count.
Deploying a Stream App

© 2012–2020, Cloudera, Inc.

Document licensed under the Creative Commons Attribution ShareAlike 4.0 License.

Cloudera.com | Documentation | Support | Community