Simple Stateful aggregations in Flink.

Prerequisites

• Hbase 2.3.3
• Hadoop 2.10.0
• Flink 1.11.2_2.11

How to use it?

A quick usage is given in below.

Flink

Flink is able to run in standalone. Start the flink cluster.

$ cd $FLINK_HOME

$ ./bin/start-cluster.sh

Upload your jar from Flink UI -> localhost:8081

Socket

Create a socket as a data source and submit your job in Flink UI.

To send 1000 record per second
$ python3 socket_data.py 1000

Output

Query generated results from hbase shell.

##Id,TotalCount,Min_Trx,Max_Trx,Avg_Trx,First_trx_time,Last_trx_time
$hbase shell
> get 'customer_habit','[trx_id]'