Fixed-Width Data Source for Apache Spark

A library for parsing FixedWidth data with Apache Spark. FixedWidth file is a flat file where each column has a fixed width (number of characters) and this is specified in a schema.

Usage

You can link against this library in your program at the following coordinates:

Scala 2.11

groupId: za.co.absa
artifactId: fixed-width_2.11
version: 0.2.0

Scala 2.12

groupId: za.co.absa
artifactId: fixed-width_2.12
version: 0.2.0

Using with Spark shell

This package can be added to Spark using the --packages command line option. For example, to include it when starting the spark shell:

Spark compiled with Scala 2.11

$SPARK_HOME/bin/spark-shell --packages za.co.absa:fixed-width_2.11:0.2.0

Spark compiled with Scala 2.12

$SPARK_HOME/bin/spark-shell --packages za.co.absa:fixed-width_2.12:0.2.0

Usage in code

import org.apache.spark.sql.types._
import org.apache.spark.sql.SparkSession

val sparkBuilder = SparkSession.builder().appName("Example")
val spark = sparkBuilder.getOrCreate()

val metadata = new MetadataBuilder().putLong("width", 10).build()
val schema = StructType(
  List(
    StructField("someColumn", StringType, true, metadata),
    StructField("secondColumn", StringType, true, metadata)
  )
)

val dataframe = spark
  .read
  .format("fixed-width")
  .option("trimValues", "true")
  .schema(schema)
  .load("/path/to/data/fixedWidthData")

Options

Option Value Explanation
trimValues true/false Should the whitespaces around data be trimmed
charset charset name (e.g. UTF-8) Any valid charset used to write the FixedWidth file

How to generate Code coverage report

sbt jacoco

Code coverage will be generated on path:

{project-root}/target/scala-{scala_version}/jacoco/report/html