This project demonstrates how to process time-series data using Apache Crunch using the simple example of sequencing trades for each stock by time.
The compiled program can be run on a Hadoop cluster with:
hadoop jar target/tradesequence-0.0.1-SNAPSHOT-job.jar /your/hdfs/input/directory /your/hdfs/output/directory
A small test data JSON file is provided in src/main/avro
. On a CDH5 cluster it can be converted to an Avro file using src/main/avro/create_test_avro.sh
. On another Hadoop distribution you can alter the script to point to your avro-tools location. The Avro data file can be used as the input for the job.