Pavan Aswathati

Apache Spark provides a lot of valuable tools for data science Data scientists use data exploration and visualization to help frame the question and fine tune the learning. Apache Zeppelin helps with this. Based on the concept of an interpreter that can be bound to any language or data processing backend, Zeppelin is a web based notebook server. As one of its backends, Zeppelin implements Spark, and other implementations, such as Hive, Markdown, D3 etc., are also available. One of the powerful features of Zeppelin Notebook is that you can view the result set of the previous section within the same framework. Zeppelin’s display system plugs into standard output. Any string that is outputted to standard output via println can be intercepted by Zeppelin’s display system if it is followed first by the interpreter command say %table, or %img, or %html etc. In our case, we would like to output the count of logs by log level as a table, so we use the following snippet of code: import org.apache.spark.sql.Row val result=sqlContext.sql("SELECT level, COUNT(1) from ambari group by level").map { case Row(level: String, count: Long) => { level + "t" + count } }.collect() This assembles the output of the groupby into a format that is suitable for the table interpreter to render. %table requires the rows each to be separated by “n” (next line) and columns to be separated by “t” (tab) characters respectively, as below: println("%table Log LeveltCountn" + result.mkString("n"))