Skip to content

Latest commit

 

History

History
14 lines (9 loc) · 686 Bytes

README.md

File metadata and controls

14 lines (9 loc) · 686 Bytes

spark_input_lib

A useful tool for reading data with Spark.

It contains it.mgaido.spark.io.FileWithHeaderReader class which allows to read files withan header ignoring it. It requires to set properly it.mgaido.spark.io.InputFileWithHeaderReader.HEADER_NUMBER_OF_LINES property in Hadoop job configuration with the number of lines each file contains as header.

This class doesn't work fine if the header is spread over multiple block, but this should not happen.... In such a case only the lines in the first block are discarded.

Scala object it.mgaido.spark.io.IOHelper performs the same operation reading a path and returning a RDD of strings without the header lines.