Pain free Spark/Cobol files integration.

Seamlessly query your COBOL/EBCDIC binary files as Spark Dataframes and streams.

Add mainframe as a source to your data engineering strategy.

Among the motivations for this project, it is possible to highlight:

• Lack of expertise in the Cobol ecosystem, which makes it hard to integrate mainframes into data engineering strategies

• Lack of support from the open-source community to initiatives in this field

• The overwhelming majority (if not all) of tools to cope with this domain are proprietary

• Several institutions struggle daily to maintain their legacy mainframes, which prevents them from evolving to more modern approaches to data management

• Mainframe data can only take part in data science activities through very expensive investments

• Supports primitive types (although some are "Cobol compiler specific")

• Supports REDEFINES, OCCURS and DEPENDING ON fields (e.g. unchecked unions and variable-size arrays)

• Supports nested structures and arrays

• Supports HDFS as well as local file systems

• The COBOL copybooks parser doesn't have a Spark dependency and can be reused for integrating into other data processing engines

We have presented Cobrix at DataWorks Summit 2019 and Spark Summit 2019 conferences. The screencasts are available here:

DataWorks Summit 2019 (General Cobrix workflow for hierarchical databases): https://www.youtube.com/watch?v=o_up7X3ZL24

Spark Summit 2019 (More detailed overview of performance optimizations): https://www.youtube.com/watch?v=BOBIdGf3Tm0

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

Scala 2.11	Scala 2.12	Scala 2.13
groupId: za.co.absa.cobrix artifactId: spark-cobol_2.11 version: 2.7.10	groupId: za.co.absa.cobrix artifactId: spark-cobol_2.12 version: 2.7.10	groupId: za.co.absa.cobrix artifactId: spark-cobol_2.13 version: 2.7.10

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

$SPARK_HOME/bin/spark-shell --packages za.co.absa.cobrix:spark-cobol_2.11:2.7.10

$SPARK_HOME/bin/spark-shell --packages za.co.absa.cobrix:spark-cobol_2.12:2.7.10

$SPARK_HOME/bin/spark-shell --packages za.co.absa.cobrix:spark-cobol_2.13:2.7.10

This repository contains several standalone example applications in examples/spark-cobol-app directory. It is a Maven project that contains several examples:

• SparkTypesApp is an example of a very simple mainframe file processing. It is a fixed record length raw data file with a corresponding copybook. The copybook contains examples of various numeric data types Cobrix supports.
• SparkCobolApp is an example of a Spark Job for handling multisegment variable record length mainframe files.
• SparkCodecApp is an example usage of a custom record header parser. This application reads a variable record length file having non-standard RDW headers. In this example RDH header is 5 bytes instead of 4
• SparkCobolHierarchical is an example processing of an EBCDIC multisegment file extracted from a hierarchical database.

The example project can be used as a template for creating Spark Application. Refer to README.md of that project for the detailed guide how to run the examples locally and on a cluster.

When running mvn clean package in examples/spark-cobol-app an uber jar will be created. It can be used to run jobs via spark-submit or spark-shell.

sbt ++{scala_version} jacoco

Code coverage will be generated on path:

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

1. Create a Spark SQLContext

2. Start a sqlContext.read operation specifying za.co.absa.cobrix.spark.cobol.source as the format

3. Inform the path to the copybook describing the files through ... .option("copybook", "path_to_copybook_file"). By default the copybook is expected to be in HDFS. You can specify that a copybook is located in the local file system by adding file:// prefix. For example, you can specify a local file like this .option("copybook", "file:///home/user/data/compybook.cpy"). Alternatively, instead of providing a path to a copybook file you can provide the contents of the copybook itself by using .option("copybook_contents", "...copybook contents...").

4. Inform the path to the HDFS directory containing the files: ... .load("path_to_directory_containing_the_binary_files")

5. Inform the query you would like to run on the Cobol Dataframe

Below is an example whose full version can be found at za.co.absa.cobrix.spark.cobol.examples.SampleApp and za.co.absa.cobrix.spark.cobol.examples.CobolSparkExample

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

val cobolDataframe = spark
  .read
  .format("cobol")
  .option("copybook", "data/test1_copybook.cob")
  .load("data/test2_data")

cobolDataframe
    .filter("RECORD.ID % 2 = 0") // filter the even values of the nested field 'RECORD_LENGTH'
    .take(10)
    .foreach(v => println(v))

The full example is available here

In some scenarios Spark is unable to find "cobol" data source by it's short name. In that case you can use the full path to the source class instead: .format("za.co.absa.cobrix.spark.cobol.source")

Cobrix assumes input data is encoded in EBCDIC. You can load ASCII files as well by specifying the following option: .option("encoding", "ascii").

If the input file is a text file (CRLF / LF are used to split records), use .option("is_text", "true").

Multisegment ASCII text files are supported using this option: .option("record_format", "D").

Cobrix has better handling of special characters and partial records using its extension format: .option("record_format", "D2").

Read more on record formats at https://www.ibm.com/docs/en/zos/2.4.0?topic=files-selecting-record-formats-non-vsam-data-sets

1. Create a Spark StreamContext

2. Import the binary files/stream conversion manager: za.co.absa.spark.cobol.source.streaming.CobolStreamer._

3. Read the binary files contained in the path informed in the creation of the SparkSession as a stream: ... streamingContext.cobolStream()

4. Apply queries on the stream: ... stream.filter("some_filter") ...

5. Start the streaming job.

Below is an example whose full version can be found at za.co.absa.cobrix.spark.cobol.examples.StreamingExample

val spark = SparkSession
  .builder()
  .appName("CobolParser")
  .master("local[2]")
  .config("duration", 2)
  .config("copybook", "path_to_the_copybook")
  .config("path", "path_to_source_directory") // could be both, local or HDFS
  .getOrCreate()          
      
val streamingContext = new StreamingContext(spark.sparkContext, Seconds(3))         
    
import za.co.absa.spark.cobol.source.streaming.CobolStreamer._ // imports the Cobol streams manager

val stream = streamingContext.cobolStream() // streams the binary files into the application    

stream
    .filter(row => row.getAs[Integer]("NUMERIC_FLD") % 2 == 0) // filters the even values of the nested field 'NUMERIC_FLD'
    .print(10)		

streamingContext.start()
streamingContext.awaitTermination()

To query mainframe files interactively using spark-shell you need to provide jar(s) containing Corbrix and it's dependencies. This can be done either by downloading all the dependencies as separate jars or by creating an uber jar that contains all of the dependencies.

Cobrix's spark-cobol data source depends on the COBOL parser that is a part of Cobrix itself and on scodec libraries to decode various binary formats.

The jars that you need to get are:

• spark-cobol_2.12-2.7.10.jar
• cobol-parser_2.12-2.7.10.jar
• scodec-core_2.12-1.10.3.jar
• scodec-bits_2.12-1.1.4.jar

Versions older than 2.7.1 also need antlr4-runtime-4.8.jar.

After that you can specify these jars in spark-shell command line. Here is an example:

$ spark-shell --packages za.co.absa.cobrix:spark-cobol_2.12:2.7.10
or 
$ spark-shell --master yarn --deploy-mode client --driver-cores 4 --driver-memory 4G --jars spark-cobol_2.12-2.7.10.jar,cobol-parser_2.12-2.7.10.jar,scodec-core_2.12-1.10.3.jar,scodec-bits_2.12-1.1.4.jar

Setting default log level to "WARN".
To adjust logging level use sc.setLogLevel(newLevel). For SparkR, use setLogLevel(newLevel).
Spark context available as 'sc' (master = yarn, app id = application_1535701365011_2721).
Spark session available as 'spark'.
Welcome to
      ____              __
     / __/__  ___ _____/ /__
    _\ \/ _ \/ _ `/ __/  '_/
   /___/ .__/\_,_/_/ /_/\_\   version 2.4.5
      /_/

Using Scala version 2.11.8 (OpenJDK 64-Bit Server VM, Java 1.8.0_171)
Type in expressions to have them evaluated.
Type :help for more information.

scala> val df = spark.read.format("cobol").option("copybook", "/data/example1/test3_copybook.cob").load("/data/example1/data")
df: org.apache.spark.sql.DataFrame = [TRANSDATA: struct<CURRENCY: string, SIGNATURE: string ... 4 more fields>]

scala> df.show(false)
+----------------------------------------------------+
|TRANSDATA                                           |
+----------------------------------------------------+
|[GBP,S9276511,Beierbauh.,0123330087,1,89341.00]     |
|[ZAR,S9276511,Etqutsa Inc.,0039003991,1,2634633.00] |
|[USD,S9276511,Beierbauh.,0038903321,0,75.71]        |
|[ZAR,S9276511,Beierbauh.,0123330087,0,215.39]       |
|[ZAR,S9276511,Test Bank,0092317899,1,643.94]        |
|[ZAR,S9276511,Xingzhoug,8822278911,1,998.03]        |
|[USD,S9276511,Beierbauh.,0123330087,1,848.88]       |
|[USD,S9276511,Beierbauh.,0123330087,0,664.11]       |
|[ZAR,S9276511,Beierbauh.,0123330087,1,55262.00]     |
+----------------------------------------------------+
only showing top 20 rows

scala>

Gathering all dependencies manually maybe a tiresome task. A better approach would be to create a jar file that contains all required dependencies (an uber jar aka fat jar).

Creating an uber jar for Cobrix is very easy. Steps to build:

• Install JDK 8
• Install SBT
• Clone Cobrix repository
• Run sbt assembly in the root directory of the repository specifying the Scala and Spark version you want to build for:

  # For Scala 2.11
  sbt -DSPARK_VERSION="2.4.8" ++2.11.12 assembly
  
  # For Scala 2.12
  sbt -DSPARK_VERSION="2.4.8" ++2.12.20 assembly
  sbt -DSPARK_VERSION="3.1.3" ++2.12.20 assembly
  sbt -DSPARK_VERSION="3.2.3" ++2.12.20 assembly
  sbt -DSPARK_VERSION="3.3.2" ++2.12.20 assembly
  sbt -DSPARK_VERSION="3.4.0" ++2.12.20 assembly
  
  # For Scala 2.13
  sbt -DSPARK_VERSION="3.3.2" ++2.13.15 assembly
  sbt -DSPARK_VERSION="3.4.0" ++2.13.15 assembly

You can collect the uber jar of spark-cobol either at spark-cobol/target/scala-2.11/ or in spark-cobol/target/scala-2.12/ depending on the Scala version you used. The fat jar will have '-bundle' suffix. You can also download pre-built bundles from https://github.com/AbsaOSS/cobrix/releases/tag/v2.7.3

Then, run spark-shell or spark-submit adding the fat jar as the option.

$ spark-shell --jars spark-cobol_2.12_3.3-2.7.11-SNAPSHOT-bundle.jar

A note for building and running tests on Windows

• java.lang.UnsatisfiedLinkError: org.apache.hadoop.io.nativeio.NativeIO$POSIX.stat is a Hadoop compatibility with Windows issue. The workaround is described here: https://stackoverflow.com/questions/41851066/exception-in-thread-main-java-lang-unsatisfiedlinkerror-org-apache-hadoop-io
• When running assembly with -DSPARK_VERSION=... on Windows, it seems an sbt compatibility with Windows issue: https://stackoverflow.com/questions/59144913/run-sbt-1-2-8-project-with-java-d-options-on-windows You can work around it by using default Spark version for a given Scala version:

  sbt ++2.11.12 assembly
  sbt ++2.12.20 assembly
  sbt ++2.13.15 assembly

Currently, specifying multiple paths in load() is not supported. Use the following syntax:

    spark
      .read
      .format("cobol")
      .option("copybook_contents", copybook)
      .option("paths", inputPaths.mkString(","))
      .load()

This library also provides convenient methods to extract Spark SQL schemas and Cobol layouts from copybooks.

If you want to extract a Spark SQL schema from a copybook by providing same options you provide to Spark:

// Same options that you use for spark.read.format("cobol").option()
val options = Map("schema_retention_policy" -> "keep_original")

val cobolSchema = CobolSchema.fromSparkOptions(Seq(copybook), options)
val sparkSchema = cobolSchema.getSparkSchema.toString()

println(sparkSchema)

If you want to extract a Spark SQL schema from a copybook using the Cobol parser directly:

import za.co.absa.cobrix.cobol.parser.CopybookParser
import za.co.absa.cobrix.cobol.reader.policies.SchemaRetentionPolicy
import za.co.absa.cobrix.spark.cobol.schema.CobolSchema

val parsedSchema = CopybookParser.parseSimple(copyBookContents)
val cobolSchema = new CobolSchema(parsedSchema, SchemaRetentionPolicy.CollapseRoot, inputFileNameField = "", generateRecordId = false)
val sparkSchema = cobolSchema.getSparkSchema.toString()

println(sparkSchema)

If you want to check the layout of the copybook:

import za.co.absa.cobrix.cobol.parser.CopybookParser

val copyBook = CopybookParser.parseSimple(copyBookContents)
println(copyBook.generateRecordLayoutPositions())

When a copybook is converted to a Spark schema, some information is lost, such as length of string fields or minimum and maximum number of elements in arrays. To preserve this information, Cobrix adds metadata to Spark schema fields. The following metadata is added:

You can access the metadata in the usual way:

// This example returns the maximum length of a string field that is the first field of the copybook
df.schema.fields(0).metadata.getLong("maxLength")

Cobrix assumes files has fixed length (F) record format by default. The record length is determined by the length of the record defined by the copybook. But you can specify the record length explicitly:

.option("record_format", "F")
.option("record_length", "250")

Fixed block record formats (FB) are also supported. The support is experimental, if you find any issues, please let us know. When the record format is 'FB' you can specify block length or number of records per block. As with 'F' if record_length is not specified, it will be determined from the copybook.

Records that have BDWs, but not rdws can be read like this:

.option("record_format", "FB")
.option("record_length", "250")

or simply

.option("record_format", "FB")

Records that have neither BDWs nor RDWs can be read like this:

.option("record_format", "FB")
.option("record_length", "250")
.option("block_length", "500")

or

.option("record_format", "FB")
.option("record_length", "250")
.option("records_per_block", "2")

More on fixed-length record formats: https://www.ibm.com/docs/en/zos/2.3.0?topic=sets-fixed-length-record-formats

Cobrix supports variable record length files. The only requirement is that such a file should contain a standard 4 byte record header known as Record Descriptor Word (RDW). Such headers are created automatically when a variable record length file is copied from a mainframe. Another type of files are variable blocked length. Such files contain Block Descriptor Word (BDW), as well as Record Descriptor Word (RDW) headers. Any such header can be either big-endian or little-endian. Also, quite often BDW headers need to be adjusted in order to be read properly. See the use cases section below.

To load variable length record file the following option should be specified:

.option("record_format", "V")

To load variable blocked length record file the following option should be specified:

.option("record_format", "VB")

More on record formats: https://www.ibm.com/docs/en/zos/2.3.0?topic=files-selecting-record-formats-non-vsam-data-sets

The space used by the headers (both BDW and RDW) should not be mentioned in the copybook if this option is used. Please refer to the 'Record headers support' section below.

If a record of the copybook contains record lengths for each record you can use 'record_length_field' like this:

.option("record_format", "F")
.option("record_length_field", "RECORD_LENGTH")

You can use expressions as well:

.option("record_format", "F")
.option("record_length_field", "RECORD_LENGTH + 10")

or

.option("record_format", "F")
.option("record_length_field", "FIELD1 * 10 + 200")

If the record field contains a string that can be mapped to a record size, you can add the mapping as a JSON:

.option("record_format", "F")
.option("record_length_field", "FIELD_STR")
.option("record_length_map", """{"SEG1":100,"SEG2":200}""")  

In order to understand the file format it is often sufficient to look at the first 4 bytes of the file (un case of RDW only files), or the first 8 bytes of a file + lookup the offset of the block (in case of BDW + RDW)

In order to determine if an RDW is a big- or little-endian, take a look at the first 4 bytes. If the first 2 bytes are zeros, it's a little-endian RDW header, otherwise it is a big-endian RDW header.

It is harder to determine if a BDW header is big- or little-endian since BDW header bytes can be all non-zero. But for VB format RDWs follow BDWs and endiness. You can determine the endiness from an RDW, and use the same option for BDW.

Mainframe data often contain only one root GROUP. In such cases such a GROUP can be considered something similar to XML rowtag. Cobrix allows either to collapse or to retain the GROUP. To turn this on use the following option:

.option("schema_retention_policy", "collapse_root")

or

.option("schema_retention_policy", "keep_original")

Let's look at an example. Let's say we have a copybook that looks like this:

       01  RECORD.
           05  ID                        PIC S9(4)  COMP.
           05  COMPANY.
               10  SHORT-NAME            PIC X(10).
               10  COMPANY-ID-NUM        PIC 9(5) COMP-3.

When "schema_retention_policy" is set to "collapse_root" (default), the root group will be collapsed and the schema will look like this (note the RECORD field is not part of the schema):

root
 |-- ID: integer (nullable = true)
 |-- COMPANY: struct (nullable = true)
 |    |-- SHORT_NAME: string (nullable = true)
 |    |-- COMPANY_ID_NUM: integer (nullable = true)

But when "schema_retention_policy" is set to "keep_original", the schema will look like this (note the RECORD field is part of the schema):

root
 |-- RECORD: struct (nullable = true)
 |    |-- ID: integer (nullable = true)
 |    |-- COMPANY: struct (nullable = true)
 |    |    |-- SHORT_NAME: string (nullable = true)
 |    |    |-- COMPANY_ID_NUM: integer (nullable = true)

You can experiment with this feature using built-in example in za.co.absa.cobrix.spark.cobol.examples.CobolSparkExample

For data that has record order dependency generation of "File_Id", "Record_Id", and "Record_Byte_Length" fields is supported. The values of the File_Id column will be unique for each file when a directory is specified as the source for data. The values of the Record_Id column will be unique and sequential record identifiers within the file.

Turn this feature on use

.option("generate_record_id", true)

The following fields will be added to the top of the schema:

root
 |-- File_Id: integer (nullable = false)
 |-- Record_Id: long (nullable = false)
 |-- Record_Byte_Length: integer (nullable = false)

You can use this option to generate raw bytes of each record as a binary field:

.option("generate_record_bytes", "true")

The following fields will be added to the top of the schema:

root
 |-- Record_Bytes: binary (nullable = false)

Variable-length records depend on headers to have their length calculated, which makes it hard to achieve parallelism while parsing.

Cobrix strives to overcome this drawback by performing a two-stages parsing. The first stage traverses the records retrieving their lengths and offsets into structures called indexes. Then, the indexes are distributed across the cluster, which allows for parallel variable-length records parsing.

However effective, this strategy may also suffer from excessive shuffling, since indexes may be sent to executors far from the actual data.

The latter issue is overcome by extracting the preferred locations for each index directly from HDFS, and then passing those locations to Spark during the creation of the RDD that distributes the indexes.

When processing large collections, the overhead of collecting the locations is offset by the benefits of locality, thus, this feature is enabled by default, but can be disabled by the configuration below:

.option("improve_locality", false)

When dealing with variable-length records, Cobrix strives to maximize locality by identifying the preferred locations in the cluster to parse each record, i.e. the nodes where the record resides.

This feature is implemented by querying HDFS about the locations of the blocks containing each record and instructing Spark to create the partition for that record in one of those locations.

However, sometimes, new nodes can be added to the cluster after the Cobol file is stored, in which case those nodes would be ignored when processing the file since they do not contain any record.

To overcome this issue, Cobrix also strives to re-balance the records among the new nodes at parsing time, as an attempt to maximize the utilization of the cluster. This is done through identifying the busiest nodes and sharing part of their burden with the new ones.

Since this is not an issue present in most cluster configurations, this feature is disabled by default, and can be enabled from the configuration below:

.option("optimize_allocation", true)

If however the option improve_locality is disabled, this option will also be disabled regardless of the value in optimize_allocation.

As you may already know a file in the mainframe world does not mean the same as in the PC world. On PCs we think of a file as a stream of bytes that we can open, read/write and close. On mainframes a file can be a set of records that we can query. Record is a blob of bytes, can have different size. Mainframe's 'filesystem' handles the mapping between logical records and physical location of data.

Details are available at this Wikipedia article (look for MVS filesystem).

So usually a file cannot simply be 'copied' from a mainframe. When files are transferred using tools like XCOM each record is prepended with an additional record header or RDW. This header allows readers of a file in PC to restore the 'set of records' nature of the file.

Mainframe files coming from IMS and copied through specialized tools contain records (the payload) having schema of DBs copybook warped with DB export tool headers wrapped with record headers. Like this:

RECORD_HEADERS ( TOOL_HEADERS ( PAYLOAD ) )

Similar to Internet's TCP protocol IP_HEADERS ( TCP_HEADERS ( PAYLOAD ) ).

TOOL_HEADERS are application dependent. Often it contains the length of the payload. But this length is sometime not very reliable. RECORD_HEADERS contain the record length (including TOOL_HEADERS length) and are proved to be reliable.

For fixed record length files record headers can be ignored since we already know the record length. But for variable record length files and for multisegment files record headers can be considered the most reliable single point of truth about record length.

You can instruct the reader to use 4 byte record headers to extract records from a mainframe file.

.option("record_format", "V")

This is very helpful for multisegment files when segments have different lengths. Since each segment has it's own copybook it is very convenient to extract segments one by one by combining record_format = V option with segment filter option.

.option("segment_field", "SEG-ID")
.option("segment_filter", "1122334")

In this example it is expected that the copybook has a field with the name 'SEG-ID'. The data source will read all segments, but will parse only ones that have SEG-ID = "1122334".

If you want to parse multiple segments, set the option 'segment_filter' to a comma separated list of the segment values. For example:

.option("segment_field", "SEG-ID")
.option("segment_filter", "1122334,1122335")

will only parse the records with SEG-ID = "1122334" OR SEG-ID = "1122335"

Custom record extractors can be used for customizing splitting of input files into a set of records. Cobrix supports text files, fixed length binary files and binary files with RDWs. If your input file is not in one of the supported formats you can implement a custom record extractor interface and provide it to spark-cobol as a option:

.option("record_extractor", "com.example.record.header.parser")

A custom record extractor needs to be a class having this precise constructor signature:

class TextRecordExtractor(ctx: RawRecordExtractorParameters) extends Serializable with RawRecordExtractor {
                             // Your implementation
                          }

A record extractor is essentially iterator of records. Each returned record is an array of bytes parsable by the copybook.

A record extractor is invoked two times. First, it is invoked at the beginning each file to go thought the file and create a sparse index. The second time it is invoked by parallel processes starting from different records in the file. The starting record number is provided in constructor. The starting file offset is available from inputStream.

RawRecordContext consists of the following fields that the custom record extractor will get from Cobrix in runtime:

• startingRecordNumber - A record number the input stream is pointing to.
• inputStream - The input stream of bytes of the input file.
• copybook - The parsed copybook of the input stream.
• additionalInfo - An arbitrary info that can be passed as an option (see below).

If your record extractor needs additional information in order to extract records properly, you can provide an arbitrary additional info to the record extracted at runtime by specifying this option:

Take a look at CustomRecordExtractorMock inside spark-cobol project to see how a custom record extractor can be built.

.option("re_additional_info", "some info")

Custom record header parsers are deprecated. Use custom record extractors instead. They are more flexible and easier to use.

If your variable length file does not have RDW headers, but has fields that can be used for determining record lengths you can provide a custom record header parser that takes starting bytes of each record and returns record lengths. In order to do that you need to create a class inheriting RecordHeaderParser and Serializable traits and provide a fully qualified class name to the following option:

.option("record_header_parser", "com.example.record.header.parser")

Cobrix provides helper methods to convert RDD[String] or RDD[Array[Byte]] to DataFrame using a copybook. This can be used if you want to use a custom logic to split the input file into records as either ASCII strings or arrays of bytes, and then parse each record using a copybook.

An example of RDD[Array[Byte]]:

import za.co.absa.cobrix.spark.cobol.Cobrix

val rdd = ???
val df = Cobrix.fromRdd
    .copybookContents(copybook)
    .option("encoding", "ebcdic") // any supported option 
    .load(rdd)

An example of ASCII Strings RDD[String]:

import za.co.absa.cobrix.spark.cobol.Cobrix

val rdd = ???
val df = Cobrix.fromRdd
    .copybookContents(copybook)
    .option("variable_size_occurs", "true") // any supported option 
    .loadText(rdd)

When converting from an RDD some of the options like record_format or generate_record_id cannot be used since the data is assumed to be already split by records and the information about file names and relative order of records is not available.

The following code pages are supported:

• common - (default) EBCDIC common characters
• common_extended - EBCDIC common characters with special characters extension
• cp037 - IBM EBCDIC US-Canada
• cp037_extended - IBM EBCDIC US-Canada with special characters extension
• cp300 - IBM EBCDIC Japanese Extended (2 byte code page)
• cp838 - IBM EBCDIC Thailand
• cp870 - IBM EBCDIC Multilingual Latin-2
• cp875 - IBM EBCDIC Greek
• cp1025 - IBM EBCDIC Multilingual Cyrillic
• cp1047 - IBM EBCDIC Latin-1/Open System
• cp1364 - (experimental support) IBM EBCDIC Korean (2 byte code page)
• cp1388 - (experimental support) IBM EBCDIC Simplified Chinese (2 byte code page)

By default, Cobrix uses common EBCDIC code page which contains only basic latin characters, numbers, and punctuation. You can specify the code page to use for all string fields by setting the ebcdic_code_page option to one of the following values:

.option("ebcdic_code_page", "cp037")

For multi-codepage files, you can specify the code page to use for each field by setting the field_code_page:<code page> option

.option("ebcdic_code_page", "cp037")
.option("field_code_page:cp1256" -> "FIELD1")
.option("field_code_page:us-ascii" -> "FIELD-2, FIELD_3")

Cobrix is primarily designed to read binary files, but you can directly use some internal functions to read ASCII text files. In ASCII text files, records are separated with newlines.

Working example 1:

    // The recommended way
    val df = spark
      .read
      .format("cobol")
      .option("copybook_contents", copybook)
      .option("ascii_charset", "ISO-8859-1") // You can choose a charset, UTF-8 is used by default
      .option("record_format", "D")
      .load(tmpFileName)

Working example 2 - Using RDDs and helper methods:

    // This is the way if you have data converted to an RDD[String] already.
    // You have full control on reading the input data records and converting them to `java.lang.String`.
    val df = Cobrix.fromRdd
        .copybookContents(copybook)
        .option("variable_size_occurs", "true") // any supported option 
        .loadText(rdd)

Working example 3 - Using RDDs and record parsers directly:

    // This is the most verbose way - creating dataframes from RDDs. But it gives full control on how text files are
    // processed before parsing actual records
    val spark = SparkSession
      .builder()
      .appName("Spark-Cobol ASCII text file")
      .master("local[*]")
      .getOrCreate()

    val copybook =
      """       01  COMPANY-DETAILS.
        |            05  SEGMENT-ID		PIC 9(1).
        |            05  STATIC-DETAILS.
        |               10  NAME      	PIC X(2).
        |
        |            05  CONTACTS REDEFINES STATIC-DETAILS.
        |               10  PERSON    	PIC X(3).
      """.stripMargin

    val parsedCopybook = CopybookParser.parse(copybook, dataEnncoding = ASCII, stringTrimmingPolicy = StringTrimmingPolicy.TrimNone)
    val cobolSchema = new CobolSchema(parsedCopybook, SchemaRetentionPolicy.CollapseRoot, "", false)
    val sparkSchema = cobolSchema.getSparkSchema

    val rddText = spark.sparkContext.textFile("src/main/resources/mini.txt")

    val recordHandler = new RowHandler()

    val rddRow = rddText
      .filter(str => str.length > 0)
      .map(str => {
        val record = RecordExtractors.extractRecord[GenericRow](parsedCopybook.ast,
          str.getBytes(),
          0,
          SchemaRetentionPolicy.CollapseRoot, handler = recordHandler)
        Row.fromSeq(record)
      })

    val dfOut = spark.createDataFrame(rddRow, sparkSchema)

    dfOut.printSchema()
    dfOut.show()

Corresponding data sample in mini.txt:

1BB
2CCC

Output:

root
 |-- SEGMENT_ID: integer (nullable = true)
 |-- STATIC_DETAILS: struct (nullable = true)
 |    |-- NAME: string (nullable = true)
 |-- CONTACTS: struct (nullable = true)
 |    |-- PERSON: string (nullable = true)

 ...

 +----------+--------------+--------+
 |SEGMENT_ID|STATIC_DETAILS|CONTACTS|
 +----------+--------------+--------+
 |         1|          [BB]|  [null]|
 |         2|          [CC]|   [CCC]|
 +----------+--------------+--------+

There, Cobrix loaded all redefines for every record. Each record contains data from all of the segments. But only one redefine is valid for every segment. Filtering is described in the following section.

When reading a multisegment file you can use Spark to clean up redefines that do not match segment ids. Cobrix will parse every redefined field for each segment. To increase performance you can specify which redefine corresponds to which segment id. This way Cobrix will parse only relevant segment redefined fields and leave the rest of the redefined fields null.

  .option("redefine-segment-id-map:0", "REDEFINED_FIELD1 => SegmentId1,SegmentId2,...")
  .option("redefine-segment-id-map:1", "REDEFINED_FIELD2 => SegmentId10,SegmentId11,...")

For the above example the load options will lok like this (last 2 options):

val df = spark
  .read
  .format("cobol")
  .option("copybook_contents", copybook)
  .option("record_format", "V")
  .option("segment_field", "SEGMENT_ID")
  .option("segment_id_level0", "C")
  .option("segment_id_level1", "P")
  .option("redefine_segment_id_map:0", "STATIC-DETAILS => C")
  .option("redefine_segment_id_map:1", "CONTACTS => P")
  .load("examples/multisegment_data/COMP.DETAILS.SEP30.DATA.dat")

The filtered data will look like this:

df.show(10)
+----------+----------+--------------------+--------------------+
|SEGMENT_ID|COMPANY_ID|      STATIC_DETAILS|            CONTACTS|
+----------+----------+--------------------+--------------------+
|         C|9377942526|[Joan Q & Z,10 Sa...|                    |
|         P|9377942526|                    |[+(277) 944 44 55...|
|         C|3483483977|[Robotrd Inc.,2 P...|                    |
|         P|3483483977|                    |[+(174) 970 97 54...|
|         P|3483483977|                    |[+(848) 832 61 68...|
|         P|3483483977|                    |[+(455) 184 13 39...|
|         C|7540764401|[Eqartion Inc.,87...|                    |
|         C|4413124035|[Xingzhoug,74 Qin...|                    |
|         C|9546291887|[ZjkLPj,5574, Tok...|                    |
|         P|9546291887|                    |[+(300) 252 33 17...|
+----------+----------+--------------------+--------------------+

In the above example invalid fields became null and the parsing is done faster because Cobrix does not need to process every redefine for each record.

A FILLER is an anonymous field that is usually used for reserving space for new fields in a fixed record length data. Or it is used to remove a field from a copybook without affecting compatibility.

      05  COMPANY.
          10  NAME      PIC X(15).
          10  FILLER    PIC X(5).
          10  ADDRESS   PIC X(25).
          10  FILLER    PIC X(125).

Such fields are dropped when imported into a Spark data frame by Cobrix. Some copybooks, however, have FILLER groups that contain non-filler fields. For example,

      05  FILLER.
          10  NAME      PIC X(15).
          10  ADDRESS   PIC X(25).
      05  FILLER.
          10  AMOUNT    PIC 9(10)V96.
          10  COMMENT   PIC X(40).

By default Cobrix will retain such fields, but will rename each such filler to a unique name so each each individual struct can be specified unambiguously. For example, in this case the filler groups will be renamed to FILLER_1 and FILLER_2. You can change this behaviour if you would like to drop such filler groups by providing this option:

.option("drop_group_fillers", "true")

In order to retain value FILLERs (e.g. non-group FILLERs) as well, use this option:

.option("drop_value_fillers", "false")

Let's imagine we have a multisegment file with 2 segments having parent-child relationships. Each segment has a different record type. The root record/segment contains company info, an address and a taxpayer number. The child segment contains a contact person for a company. Each company can have zero or more contact persons. So each root record can be followed by zero or more child records.

To load such data in Spark the first thing you need to do is to create a copybook that contains all segment specific fields in redefined groups. Here is the copybook for our example:

        01  COMPANY-DETAILS.
            05  SEGMENT-ID        PIC X(5).
            05  COMPANY-ID        PIC X(10).
            05  STATIC-DETAILS.
               10  COMPANY-NAME      PIC X(15).
               10  ADDRESS           PIC X(25).
               10  TAXPAYER.
                  15  TAXPAYER-TYPE  PIC X(1).
                  15  TAXPAYER-STR   PIC X(8).
                  15  TAXPAYER-NUM  REDEFINES TAXPAYER-STR
                                     PIC 9(8) COMP.

            05  CONTACTS REDEFINES STATIC-DETAILS.
               10  PHONE-NUMBER      PIC X(17).
               10  CONTACT-PERSON    PIC X(28).

The 'SEGMENT-ID' and 'COMPANY-ID' fields are present in all of the segments. The 'STATIC-DETAILS' group is present only in the root record. The 'CONTACTS' group is present only in child record. Notice that 'CONTACTS' redefine 'STATIC-DETAILS'.

Because the records have different lengths use record_format = V or record_format = VB depending of the record format.

If you load this file as is you will get the schema and the data similar to this.

val df = spark
  .read
  .format("cobol")
  .option("copybook", "/path/to/thecopybook")
  .option("record_format", "V")
  .load("examples/multisegment_data")

df.printSchema()
root
 |-- SEGMENT_ID: string (nullable = true)
 |-- COMPANY_ID: string (nullable = true)
 |-- STATIC_DETAILS: struct (nullable = true)
 |    |-- COMPANY_NAME: string (nullable = true)
 |    |-- ADDRESS: string (nullable = true)
 |    |-- TAXPAYER: struct (nullable = true)
 |    |    |-- TAXPAYER_TYPE: string (nullable = true)
 |    |    |-- TAXPAYER_STR: string (nullable = true)
 |    |    |-- TAXPAYER_NUM: integer (nullable = true)
 |-- CONTACTS: struct (nullable = true)
 |    |-- PHONE_NUMBER: string (nullable = true)
 |    |-- CONTACT_PERSON: string (nullable = true)

df.show(10)
+----------+----------+--------------------+--------------------+
|SEGMENT_ID|COMPANY_ID|      STATIC_DETAILS|            CONTACTS|
+----------+----------+--------------------+--------------------+
|         C|9377942526|[Joan Q & Z,10 Sa...|[Joan Q & Z     1...|
|         P|9377942526|[+(277) 944 44 5,...|[+(277) 944 44 55...|
|         C|3483483977|[Robotrd Inc.,2 P...|[Robotrd Inc.   2...|
|         P|3483483977|[+(174) 970 97 5,...|[+(174) 970 97 54...|
|         P|3483483977|[+(848) 832 61 6,...|[+(848) 832 61 68...|
|         P|3483483977|[+(455) 184 13 3,...|[+(455) 184 13 39...|
|         C|7540764401|[Eqartion Inc.,87...|[Eqartion Inc.  8...|
|         C|4413124035|[Xingzhoug,74 Qin...|[Xingzhoug      7...|
|         C|9546291887|[ZjkLPj,5574, Tok...|[ZjkLPj         5...|
|         P|9546291887|[+(300) 252 33 1,...|[+(300) 252 33 17...|
+----------+----------+--------------------+--------------------+

As you can see Cobrix loaded all redefines for every record. Each record contains data from all of the segments. But only one redefine is valid for every segment. So we need to split the data set into 2 datasets or tables. The distinguisher is the 'SEGMENT_ID' field. All company details will go into one data sets (segment id = 'C' [company]) while contacts will go in the second data set (segment id = 'P' [person]). While doing the split we can also collapse the groups so the table won't contain nested structures. This can be helpful to simplify the analysis of the data.

While doing it you might notice that the taxpayer number field is actually a redefine. Depending on the 'TAXPAYER_TYPE' either 'TAXPAYER_NUM' or 'TAXPAYER_STR' is used. We can resolve this in our Spark app as well.

Starting from spark-cobol version 1.1.0 hierarchical structure of multisegment records can be restored automatically. In order to do this you need to provide:

• A segment ID field that will be used to distinguish segment types.
• A segmentId to redefine fields mapping that will be used to map each segment to a redefine field.
• A parent-child relationship between segments identified by segment redefine fields.

When all of the above is specified Cobrix can reconstruct hierarchical nature of records by making child segments nested arrays of parent segments. Arbitrary levels of hierarchy and arbitrary number of segments is supported.

val df = spark
  .read
  .format("cobol")
  .option("copybook", "/path/to/thecopybook")
  .option("record_format", "V")

  // Specifies a field containing a segment id
  .option("segment_field", "SEGMENT_ID")
  
  // Specifies a mapping between segment ids and segment redefine fields
  .option("redefine_segment_id_map:1", "STATIC-DETAILS => C")
  .option("redefine-segment-id-map:2", "CONTACTS => P")
  
  // Specifies a parent-child relationship
  .option("segment-children:1", "STATIC-DETAILS => CONTACTS")
  
  .load("examples/multisegment_data")

The output schema will be

scala> df.printSchema()

root
 |-- SEGMENT_ID: string (nullable = true)
 |-- COMPANY_ID: string (nullable = true)
 |-- STATIC_DETAILS: struct (nullable = true)
 |    |-- COMPANY_NAME: string (nullable = true)
 |    |-- ADDRESS: string (nullable = true)
 |    |-- TAXPAYER: struct (nullable = true)
 |    |    |-- TAXPAYER_TYPE: string (nullable = true)
 |    |    |-- TAXPAYER_STR: string (nullable = true)
 |    |    |-- TAXPAYER_NUM: integer (nullable = true)
 |    |-- CONTACTS: array (nullable = true)
 |    |    |-- element: struct (containsNull = true)
 |    |    |    |-- PHONE_NUMBER: string (nullable = true)
 |    |    |    |-- CONTACT_PERSON: string (nullable = true)

Notice that contacts now is an array of structs. That is a company static details can contain zero or mor contacts. A possible hierarchical record output is

scala> import za.co.absa.cobrix.spark.cobol.utils.SparkUtils

scala> println(SparkUtils.prettyJSON(df.toJSON.take(1).mkString("[", ", ", "]")))
{
  "SEGMENT_ID" : "C",
  "COMPANY_ID" : "9377942526",
  "STATIC_DETAILS" : {
    "COMPANY_NAME" : "Joan Q & Z",
    "ADDRESS" : "10 Sandton, Johannesburg",
    "TAXPAYER" : {
      "TAXPAYER_TYPE" : "A",
      "TAXPAYER_STR" : "92714306",
      "TAXPAYER_NUM" : 959592241
    },
    "CONTACTS" : [ {
      "PHONE_NUMBER" : "+(174) 970 97 54",
      "CONTACT_PERSON" : "Tyesha Debow"
    }, {
      "PHONE_NUMBER" : "+(848) 832 61 68",
      "CONTACT_PERSON" : "Mindy Celestin"
    }, {
      "PHONE_NUMBER" : "+(455) 184 13 39",
      "CONTACT_PERSON" : "Mabelle Winburn"
    } ]
  }
}

An advanced hierarchical example with multiple levels of nesting and multiple segments on each level is available as a unit test za/co/absa/cobrix/spark/cobol/source/integration/Test17HierarchicalSpec.scala.

Alternatively, hierarchical record structure can be reconstructed manually by extracting each segment and joining segments together. This a is more complicated process, but it provides more control.

import spark.implicits._

val dfCompanies = df
  // Filtering the first segment by segment id
  .filter($"SEGMENT_ID"==="C")
  // Selecting fields that are only available in the first segment
  .select($"COMPANY_ID", $"STATIC_DETAILS.COMPANY_NAME", $"STATIC_DETAILS.ADDRESS",
  // Resolving the taxpayer redefine
    when($"STATIC_DETAILS.TAXPAYER.TAXPAYER_TYPE" === "A", $"STATIC_DETAILS.TAXPAYER.TAXPAYER_STR")
      .otherwise($"STATIC_DETAILS.TAXPAYER.TAXPAYER_NUM").cast(StringType).as("TAXPAYER"))

The resulting table looks like this:

dfCompanies.show(10, truncate = false)
+----------+-------------+-------------------------+--------+
|COMPANY_ID|COMPANY_NAME |ADDRESS                  |TAXPAYER|
+----------+-------------+-------------------------+--------+
|9377942526|Joan Q & Z   |10 Sandton, Johannesburg |92714306|
|3483483977|Robotrd Inc. |2 Park ave., Johannesburg|31195396|
|7540764401|Eqartion Inc.|871A Forest ave., Toronto|87432264|
|4413124035|Xingzhoug    |74 Qing ave., Beijing    |50803302|
|9546291887|ZjkLPj       |5574, Tokyo              |73538919|
|9168453994|Test Bank    |1 Garden str., London    |82573513|
|4225784815|ZjkLPj       |5574, Tokyo              |96136195|
|8463159728|Xingzhoug    |74 Qing ave., Beijing    |17785468|
|8180356010|Eqartion Inc.|871A Forest ave., Toronto|79054306|
|7107728116|Xingzhoug    |74 Qing ave., Beijing    |70899995|
+----------+-------------+-------------------------+--------+

This looks like a valid and clean table containing the list of companies. Now let's do the same for the second segment.

    val dfContacts = df
      // Filtering the second segment by segment id
      .filter($"SEGMENT_ID"==="P")
      // Selecting the fields only valid for the second segment
      .select($"COMPANY_ID", $"CONTACTS.CONTACT_PERSON", $"CONTACTS.PHONE_NUMBER")

The resulting data loons like this:

dfContacts.show(10, truncate = false)
+----------+--------------------+----------------+
|COMPANY_ID|CONTACT_PERSON      |PHONE_NUMBER    |
+----------+--------------------+----------------+
|9377942526|Janiece Newcombe    |+(277) 944 44 55|
|3483483977|Tyesha Debow        |+(174) 970 97 54|
|3483483977|Mindy Celestin      |+(848) 832 61 68|
|3483483977|Mabelle Winburn     |+(455) 184 13 39|
|9546291887|Carrie Celestin     |+(300) 252 33 17|
|9546291887|Edyth Deveau        |+(907) 101 70 64|
|9546291887|Jene Norgard        |+(694) 918 17 44|
|9168453994|Timika Bourke       |+(768) 691 44 85|
|9168453994|Lynell Riojas       |+(695) 918 33 16|
|4225784815|Jene Mackinnon      |+(540) 937 33 71|
+----------+--------------------+----------------+

This looks good as well. The table contains the list of contact persons for companies. This data set contains the 'COMPANY_ID' field which we can use later to join the tables. But often there are no such fields in data imported from hierarchical databases. If that is the case Cobrix can help you craft such fields automatically. Use 'segment_field' to specify a field that contain the segment id. Use 'segment_id_level0' to ask Cobrix to generate ids for the particular segments. We can use 'segment_id_level1' to generate child ids as well. If children records can contain children of their own we can use 'segment_id_level2' etc.

val df = spark
  .read
  .format("cobol")
  .option("copybook_contents", copybook)
  .option("record_format", "V")
  .option("segment_field", "SEGMENT_ID")
  .option("segment_id_level0", "C")
  .option("segment_id_level1", "P")
  .load("examples/multisegment_data/COMP.DETAILS.SEP30.DATA.dat")

Sometimes, the leaf level has many segments. In this case, you can use _ as the list of segment ids to specify 'the rest of segment ids', like this:

val df = spark
  .read
  .format("cobol")
  .option("copybook_contents", copybook)
  .option("record_format", "V")
  .option("segment_field", "SEGMENT_ID")
  .option("segment_id_level0", "C")
  .option("segment_id_level1", "_")
  .load("examples/multisegment_data/COMP.DETAILS.SEP30.DATA.dat")

The result of both above code snippets is the same.

The resulting table will look like this:

df.show(10)
+------------------+-----------------------+----------+----------+--------------------+--------------------+
|           Seg_Id0|                Seg_Id1|SEGMENT_ID|COMPANY_ID|      STATIC_DETAILS|            CONTACTS|
+------------------+-----------------------+----------+----------+--------------------+--------------------+
|20181219130609_0_0|                   null|         C|9377942526|[Joan Q & Z,10 Sa...|[Joan Q & Z     1...|
|20181219130609_0_0|20181219130723_0_0_L1_1|         P|9377942526|[+(277) 944 44 5,...|[+(277) 944 44 55...|
|20181219130609_0_2|                   null|         C|3483483977|[Robotrd Inc.,2 P...|[Robotrd Inc.   2...|
|20181219130609_0_2|20181219130723_0_2_L1_1|         P|3483483977|[+(174) 970 97 5,...|[+(174) 970 97 54...|
|20181219130609_0_2|20181219130723_0_2_L1_2|         P|3483483977|[+(848) 832 61 6,...|[+(848) 832 61 68...|
|20181219130609_0_2|20181219130723_0_2_L1_3|         P|3483483977|[+(455) 184 13 3,...|[+(455) 184 13 39...|
|20181219130609_0_6|                   null|         C|7540764401|[Eqartion Inc.,87...|[Eqartion Inc.  8...|
|20181219130609_0_7|                   null|         C|4413124035|[Xingzhoug,74 Qin...|[Xingzhoug      7...|
|20181219130609_0_8|                   null|         C|9546291887|[ZjkLPj,5574, Tok...|[ZjkLPj         5...|
|20181219130609_0_8|20181219130723_0_8_L1_1|         P|9546291887|[+(300) 252 33 1,...|[+(300) 252 33 17...|
+------------------+-----------------------+----------+----------+--------------------+--------------------+

The data now contain 2 additional fields: 'Seg_Id0' and 'Seg_Id1'. The 'Seg_Id0' is an autogenerated id for each root record. It is also unique for a root record. After splitting the segments you can use Seg_Id0 to join both tables. The 'Seg_Id1' field contains a unique child id. It is equal to 'null' for all root records but uniquely identifies child records.

You can now split these 2 segments and join them by Seg_Id0. The full example is available at spark-cobol/src/main/scala/za/co/absa/cobrix/spark/cobol/examples/CobolSparkExample2.scala

To run it from an IDE you'll need to change Scala and Spark dependencies from 'provided' to 'compile' so the jar file would contain all the dependencies. This is because Cobrix is a library to be used in Spark job projects. Spark jobs uber jars should not contain Scala and Spark dependencies since Hadoop clusters have their Scala and Spark dependencies provided by the infrastructure. Including Spark and Scala dependencies in an uber jar can produce binary incompatibilities when these jars are used in spark-submit and spark-shell.

Here is our example tables to join:

dfCompanies.show(10, truncate = false)
+--------------------+----------+-------------+-------------------------+--------+
|Seg_Id0             |COMPANY_ID|COMPANY_NAME |ADDRESS                  |TAXPAYER|
+--------------------+----------+-------------+-------------------------+--------+
|20181219130723_0_0  |9377942526|Joan Q & Z   |10 Sandton, Johannesburg |92714306|
|20181219130723_0_2  |3483483977|Robotrd Inc. |2 Park ave., Johannesburg|31195396|
|20181219130723_0_6  |7540764401|Eqartion Inc.|871A Forest ave., Toronto|87432264|
|20181219130723_0_7  |4413124035|Xingzhoug    |74 Qing ave., Beijing    |50803302|
|20181219130723_0_8  |9546291887|ZjkLPj       |5574, Tokyo              |73538919|
|20181219130723_0_12 |9168453994|Test Bank    |1 Garden str., London    |82573513|
|20181219130723_0_15 |4225784815|ZjkLPj       |5574, Tokyo              |96136195|
|20181219130723_0_20 |8463159728|Xingzhoug    |74 Qing ave., Beijing    |17785468|
|20181219130723_0_24 |8180356010|Eqartion Inc.|871A Forest ave., Toronto|79054306|
|20181219130723_0_27 |7107728116|Xingzhoug    |74 Qing ave., Beijing    |70899995|
+--------------------+----------+-------------+-------------------------+--------+

dfContacts.show(13, truncate = false)
+-------------------+----------+-------------------+----------------+
|Seg_Id0            |COMPANY_ID|CONTACT_PERSON     |PHONE_NUMBER    |
+-------------------+----------+-------------------+----------------+
|20181219130723_0_0 |9377942526|Janiece Newcombe    |+(277) 944 44 55|
|20181219130723_0_2 |3483483977|Tyesha Debow        |+(174) 970 97 54|
|20181219130723_0_2 |3483483977|Mindy Celestin      |+(848) 832 61 68|
|20181219130723_0_2 |3483483977|Mabelle Winburn     |+(455) 184 13 39|
|20181219130723_0_8 |9546291887|Carrie Celestin     |+(300) 252 33 17|
|20181219130723_0_8 |9546291887|Edyth Deveau        |+(907) 101 70 64|
|20181219130723_0_8 |9546291887|Jene Norgard        |+(694) 918 17 44|
|20181219130723_0_12|9168453994|Timika Bourke       |+(768) 691 44 85|
|20181219130723_0_12|9168453994|Lynell Riojas       |+(695) 918 33 16|
|20181219130723_0_15|4225784815|Jene Mackinnon      |+(540) 937 33 71|
|20181219130723_0_15|4225784815|Timika Concannon    |+(122) 216 11 25|
|20181219130723_0_15|4225784815|Jene Godfrey        |+(285) 643 50 47|
|20181219130723_0_15|4225784815|Gabriele Winburn    |+(489) 644 53 67|
+-------------------+----------+-------------------+----------------+

Let's now join these tables.

The join statement in Spark:

val dfJoined = dfCompanies.join(dfContacts, "Seg_Id0")

The joined data looks like this:

dfJoined.show(13, truncate = false)
+--------------------+----------+-------------+-------------------------+--------+----------+--------------------+----------------+
|Seg_Id0             |COMPANY_ID|COMPANY_NAME |ADDRESS                  |TAXPAYER|COMPANY_ID|CONTACT_PERSON      |PHONE_NUMBER    |
+--------------------+----------+-------------+-------------------------+--------+----------+--------------------+----------------+
|20181219130723_0_0  |9377942526|Joan Q & Z   |10 Sandton, Johannesburg |92714306|9377942526|Janiece Newcombe    |+(277) 944 44 55|
|20181219131239_0_2  |3483483977|Robotrd Inc. |2 Park ave., Johannesburg|31195396|3483483977|Mindy Celestin      |+(848) 832 61 68|
|20181219131239_0_2  |3483483977|Robotrd Inc. |2 Park ave., Johannesburg|31195396|3483483977|Tyesha Debow        |+(174) 970 97 54|
|20181219131239_0_2  |3483483977|Robotrd Inc. |2 Park ave., Johannesburg|31195396|3483483977|Mabelle Winburn     |+(455) 184 13 39|
|20181219131344_0_8  |9546291887|ZjkLPj       |5574, Tokyo              |73538919|9546291887|Jene Norgard        |+(694) 918 17 44|
|20181219131344_0_8  |9546291887|ZjkLPj       |5574, Tokyo              |73538919|9546291887|Edyth Deveau        |+(907) 101 70 64|
|20181219131344_0_8  |9546291887|ZjkLPj       |5574, Tokyo              |73538919|9546291887|Carrie Celestin     |+(300) 252 33 17|
|20181219131344_0_12 |9168453994|Test Bank    |1 Garden str., London    |82573513|9168453994|Timika Bourke       |+(768) 691 44 85|
|20181219131344_0_12 |9168453994|Test Bank    |1 Garden str., London    |82573513|9168453994|Lynell Riojas       |+(695) 918 33 16|
|20181219131344_0_15 |4225784815|ZjkLPj       |5574, Tokyo              |96136195|4225784815|Jene Mackinnon      |+(540) 937 33 71|
|20181219131344_0_15 |4225784815|ZjkLPj       |5574, Tokyo              |96136195|4225784815|Timika Concannon    |+(122) 216 11 25|
|20181219131344_0_15 |4225784815|ZjkLPj       |5574, Tokyo              |96136195|4225784815|Jene Godfrey        |+(285) 643 50 47|
|20181219131344_0_15 |4225784815|ZjkLPj       |5574, Tokyo              |96136195|4225784815|Gabriele Winburn    |+(489) 644 53 67|
+--------------------+----------+-------------+-------------------------+--------+----------+--------------------+----------------+

Again, the full example is available at spark-cobol/src/main/scala/za/co/absa/cobrix/spark/cobol/examples/CobolSparkExample2.scala

Some encoding formats are not expressible by the standard copybook spec. Cobrix has extensions to help you decode fields encoded in this way.

Loading multiple paths in the standard way is not supported.

 val df = spark
   .read
   .format("cobol")
   .option("copybook_contents", copybook)
   .load("/path1", "/paths2")

But there is a Cobrix extension that allows you to load multiple paths:

 val df = spark
   .read
   .format("cobol")
   .option("copybook_contents", copybook)
   .option("data_paths", "/path1,/paths2")
   .load()

Cobrix expects all binary numbers to be big-endian. If you have a binary number in the little-endian format, use COMP-9 (Cobrix extension) instead of COMP or COMP-5 for the affected fields.

For example, 0x01 0x02 is 1 + 2*256 = 513 in big-endian (COMP) and 1*256 + 2 = 258 (COMP-9) in little-endian.

  10 NUM  PIC S9(8) COMP.    ** Big-endian
  10 NUM  PIC S9(8) COMP-9.  ** Little-endian

Unsigned backed numbers are encoded as BCD (COMP-3) without the sign nibble. For example, bytes 0x12 0x34 encode the number 1234. As of 2.6.2 Cobrix supports decoding such numbers using an extension. Use COMP-3U for unsigned packed numbers.

The 'COMP-3U' usage

  10 NUM  PIC X(4) COMP-3U.

Note that when using X 4 refers to the number of bytes the field occupies. Here, the number of digits is 4*2 = 8.

  10 NUM  PIC 9(8) COMP-3U.

When using 9 8 refers to the number of digits the number has. Here, the size of the field in bytes is 8/2 = 4.

  10 NUM  PIC 9(6)V99 COMP-3U.

You can have decimals when using COMP-3 as well.

Flattening could be helpful when migrating data from mainframe data with fields that have OCCURs (arrays) to a relational databases that do not support nested arrays.

Cobrix has a method that can flatten the schema automatically given a DataFrame produced by spark-cobol.

Spark Scala example:

val dfFlat = SparkUtils.flattenSchema(df, useShortFieldNames = false)

PySpark example

from pyspark.sql import SparkSession, DataFrame, SQLContext
from pyspark.sql.types import StructType, StructField, StringType, IntegerType, ArrayType
from py4j.java_gateway import java_import

schema = StructType([
   StructField("id", IntegerType(), True),
   StructField("name", StringType(), True),
   StructField("subjects", ArrayType(StringType()), True)
])

# Sample data
data = [
   (1, "Alice", ["Math", "Science"]),
   (2, "Bob", ["History", "Geography"]),
   (3, "Charlie", ["English", "Math", "Physics"])
]

# Create a test DataFrame
df = spark.createDataFrame(data, schema)

# Show the Dataframe before flattening
df.show()

# Flatten the schema using Cobrix Scala 'SparkUtils.flattenSchema' method
sc = spark.sparkContext
java_import(sc._gateway.jvm, "za.co.absa.cobrix.spark.cobol.utils.SparkUtils")
dfFlatJvm = spark._jvm.SparkUtils.flattenSchema(df._jdf, False)
dfFlat = DataFrame(dfFlatJvm, SQLContext(sc))

# Show the Dataframe after flattening
dfFlat.show(truncate=False)
dfFlat.printSchema()

The output looks like this:

# Before flattening
+---+-------+------------------------+
|id |name   |subjects                |
+---+-------+------------------------+
|1  |Alice  |[Math, Science]         |
|2  |Bob    |[History, Geography]    |
|3  |Charlie|[English, Math, Physics]|
+---+-------+------------------------+

# After flattening
+---+-------+----------+----------+----------+
|id |name   |subjects_0|subjects_1|subjects_2|
+---+-------+----------+----------+----------+
|1  |Alice  |Math      |Science   |null      |
|2  |Bob    |History   |Geography |null      |
|3  |Charlie|English   |Math      |Physics   |
+---+-------+----------+----------+----------+

common_extended, cp037_extended are code pages supporting non-printable characters that converts to ASCII codes below 32.

Performance tests were performed on synthetic datasets. The setup and results are as follows.

• Spark 2.2.1
• Driver memory: 4GB
• Driver cores: 4
• Executor memory: 4GB
• Cores per executor: 1

The test Spark Application is just a conversion from the mainframe format to Parquet.

For fixed record length tests:

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

    val copybook = "...copybook contents..."
    val df = spark
      .read
      .format("cobol")
      .option("copybook_contents", copybook)
      .load(args(0))
    
      df.write.mode(SaveMode.Overwrite).parquet(args(1))

For multisegment variable lengths tests:

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

    val copybook = "...copybook contents..."
    val df = spark
      .read
      .format("cobol")
      .option("copybook_contents", copybook)
      .option("generate_record_id", true)
      .option("record_format", "V")
      .option("segment_field", "SEGMENT_ID")
      .option("segment_id_level0", "C")
      .load(args(0))
    
      df.write.mode(SaveMode.Overwrite).parquet(args(1))

• Raw (single segment) fixed length record file
• 167 columns
• 1341 bytes per record
• 30,000,000 records
• 40 GB single input file size
• The data can be generated using za.co.absa.cobrix.cobol.parser.examples.TestDataGen6TypeVariety generator app

• Multisegment variable length record file
• A copybook containing very few fields
• 68 bytes per segment 1 record and 64 bytes per segment 2 record
• 640,000,000 records
• 40 GB single input file size
• The data can be generated using za.co.absa.cobrix.cobol.parser.examples.TestDataGen3Companies generator app

• Multisegment variable length record file
• A copybook containing a segment with arrays of 2000 struct elements. This reproduces a case when a copybook contain a lot of fields.
• 16068 bytes per segment 1 record and 64 bytes per segment 2 record
• 8,000,000 records
• 40 GB single input file size
• The data can be generated using za.co.absa.cobrix.cobol.parser.examples.generatorsTestDataGen4CompaniesWide generator app

sbt jacoco

Code coverage will be generated on path:

{local-path}\fixed-width\target\scala-2.XY\jacoco\report\html

This is a new section where we are going to post common questions and workarounds from GitHub issues raised by our users.

Q: Numeric COMP or COMP-5 data is decoded incorrectly. Specifically, small values look like very big values

A: This is often a sign that the binary data is little-endian. Cobrix expects all binary data to be big-endian. The workaround is to use COMP-9 (Cobrix extension) instead of COMP and COMP-5 for the affected fields.

Q: Getting the following error when using Spark < 2.4.3:

ANTLR Tool version 4.7.2 used for code generation does not match the current runtime version 4.5.3ANTLR 
Runtime version 4.7.2 used for parser compilation does not match the current runtime version 4.5.321/12/20 11:42:54
ERROR ApplicationMaster: User class threw exception: java.lang.ExceptionInInitializerError

A: Option 1: Use Spark 2.4.3 or higher. Option 2: Use 'sbt assembly' as stated above in README to generate your spark-cobol artifact tailored for your Spark version. The artifact shades ANTLR so the incompatibility should be resolved.

Q: Getting exceptions from Hadoop classes when running Cobrix test suite on Windows:

exception or error caused a run to abort: org.apache.hadoop.io.nativeio.NativeIO$POSIX.stat(Ljava/lang/String;)Lorg/apache/hadoop/io/nativeio/NativeIO$POSIX$Stat;
java.lang.UnsatisfiedLinkError: org.apache.hadoop.io.nativeio.NativeIO$POSIX.stat(Ljava/lang/String;)Lorg/apache/hadoop/io/nativeio/NativeIO$POSIX$Stat;
at org.apache.hadoop.io.nativeio.NativeIO$POSIX.stat(Native Method)
at org.apache.hadoop.io.nativeio.NativeIO$POSIX.getStat(NativeIO.java:608)

A: Update hadoop dll to version 3.2.2 or newer.

• 2.7.10 released 19 December 2024.

  • #728 Added CP1145 code page (Spain and Latin America).
  • #731 Added an option to copy data type when copying metadata.

    SparkUtils.copyMetadata(schemaFrom, schemaTo, copyDataType = true)

• 2.7.9 released 8 November 2024.

  • #722 Added more EBCDIC code pages (mostly European).

• 2.7.8 released 21 October 2024.

  • #719 Improve error message for fixed record length files when the record size does not divide file size.

• 2.7.7 released 10 October 2024.

  • #702 Fix a race condition for fixed record length file processing. Thanks @vinodkc for the fix and @pinakigit for testing!
  • #715 Fix Jacoco report and update Scala versions.

• 2.7.6 released 26 September 2024.

  • #710 Fix index generation for files with record length fields or expressions.
  • #712 Add an option for explicitly logging layout positions (false by default).

    // Enable logging of layout positions
    .option("debug_layout_positions", "true")

• 2.7.5 released 19 August 2024.

  • #703 Add maximum length for generated segment id fields, like seg_id0, seg_id1, etc.

• 2.7.4 released 31 July 2024.

  • #697 Improve metadata merging method in Spark Utils. Add conflict resolution and merge flags

    // Field metadata can be lost during various transformations.
    // You can copy metadata from one schema to another directly
    val df1 = ??? //  A dataframe with metadata
    val df2 = ??? //  A dataframe without metadata
    val mergedSchema = SparkUtils.copyMetadata(df1.schema, df2.schema)
    
    // Create the new dataframe based on the schema with merged metadata 
    val newDf = spark.createDataFrame(df2.rdd, mergedSchema)

• 2.7.3 released 17 July 2024.

  • #678 Add the ability to generate Spark schema based on strict integral precision:

    // `decimal(n,0)` will be used instead of `integer` and `long`
    .option("strict_integral_precision", "true")

  • #689 Add support for '_' for hierarchical key generation at leaf level:

    .option("segment_id_level0", "SEG0") // Root segment
    .option("segment_id_level1", "_")    // Leaf segment (use 'all other' segment IDs)

• 2.7.2 released 7 June 2024.

  • #684 Fixed failing to read a data file in certain combination of options.
  • #685 Added methods to flatten schema of a dataframe more effective than flattenSchema(), but does not flatten arrays:

    // df - a DataFrame with nested structs
    val flatDf = SparkUtils.unstructDataFrame(df)
    // flatDf the same dataframe with all nested fields promoted to the top level.

• 2.7.1 released 4 June 2024.

  • #680 Shaded ANTLR runtime in 'cobol-parser' to avoid conflicts with various versions of Spark that uses ANTLR as well.
  • #678 Added an experimental method SparkUtils.covertIntegralToDecimal() for applying extended metadata to a DataFrame.

• 2.7.0 released 23 April 2024.

  • #666 Added support for record length value mapping.

    .option("record_format", "F")
    .option("record_length_field", "FIELD_STR")
    .option("record_length_map", """{"SEG1":100,"SEG2":200}""")

  • #669 Allow 'V' to be at the end of scaled PICs.

         10  SCALED-DECIMAL-FIELD    PIC S9PPPV      DISPLAY.

  • #672 Add the ability to parse copybooks with options normally passed to the spark-cobol Spark data source.

    // Same options that you use for spark.read.format("cobol").option()
    val options = Map("schema_retention_policy" -> "keep_original")
    
    val cobolSchema = CobolSchema.fromSparkOptions(Seq(copybook), options)
    val sparkSchema = cobolSchema.getSparkSchema.toString()
    
    println(sparkSchema)

  • #674 Extended the usage of indexes for variable record length files with a record length field.

    .option("record_length_field", "RECORD-LENGTH")
    .option("enable_indexes", "true") // true by default so can me omitted

• 2.6.11 released 8 April 2024.

  • #659 Fixed record length option when record id generation is turned on.

• 2.6.10 released 17 January 2024.

  • #653 Add support for new EBCDIC code pages: 273, 500, 1140, 1141, 1148.

• 2.6.9 released 16 October 2023.

  • #634 Retain metadata when flattening the schema in SparkUtils.
  • #644 Add support for Spark 3.5.0.

• 2.6.8 released 1 June 2023.

  • #624 Add support for binary fields that have PIC X and USAGE COMP.

• 2.6.7 released 6 May 2023.

  • #620 Fixed a regression bug that made a breaking change to custom record extractors. The source code compatibility has been restored.

• 2.6.6 released 5 May 2023.

  • #601 Fixed file_start_offset and file_end_offset options for VB record format (BDW+RDW).
  • #614 Fixed catching a state when a custom record extractor does not conform to the contract.
  • #613 Added the ability of custom record extractors to get header stream pointing to teh beginning of the file.
  • #607 Added minimum_record_length and maximum_record_length options.

• 2.6.5 released 5 April 2023.

  • #539 Fixed 'cp300', and added experimental support for 'cp1364' and 'cp1388' code pages (thanks @BenceBenedek).
  • #590 Changed from .option("extended_metadata", true) to .option("metadata", "extended") allowing other modes like 'basic' (default) and 'false' (disable metadata).
  • #593 Added option .option("generate_record_bytes", true) that adds a field containing raw bytes of each record decoded.

• 2.6.4 released 3 March 2023.

  • #576 Added the ability to create DataFrames from RDDs plus a copybook using .Cobrix.fromRdd() extension (look for 'Cobrix.fromRdd' for examples).
  • #574 Added the ability to read data files with fields encoded using multiple code pages using (.option("field_code_page:cp037", "FIELD-1,FIELD_2")).
  • #538 Added experimental support for cp00300, the 2 byte Japanese code page (thanks @BenceBenedek).

• 2.6.3 released 1 February 2023.

  • #550 Added .option("extended_metadata", true) option that adds many additional metadata fields (PIC, USAGE, etc) to the generated Spark schema.
  • #567 Added support for new code pages 838, 870, 1025 (Thanks @sree018).
  • #569 Added support for field length expressions based on filed on the copybook See Variable length records support.
  • #572 Improved performance of non-UTF8 encoded ASCII test files.
  • #560 Fixed 'segment_filter' for ASCII files when Spark test splitter is used.

• 2.6.2 released 3 January 2023.

  • #516 Added support for unsigned packed numbers via a Cobrix extension (COMP-3U).
  • #542 Added .option("filler_naming_policy", "previous_field_name") allowing for a different filler naming strategy.
  • #544 Added data_paths option to replace paths option that conflicts with Sparks internal option paths.
  • #545 Added support for string debug columns for ASCII (D/D2/T) files (.option("debug", "string")).
  • #543 Improved performance of processing ASCII text (D/D2/T) files with variable OCCURS.
  • #553 Fixed variable occurs now working properly with basic ASCII record format (D2).
  • #556 Fixed file_end_offset option dropping records from the end of partitions instead of end of files.

• 2.6.1 released 2 December 2022.

  • #531 Added support for CP1047 EBCDIC code page.
  • #532 Added Jacoco code coverage support.
  • #529 Fixed unit tests failing on Windows.

• 2.6.0 released 14 October 2022.

  • #514 Added support for Scala 2.13 and Spark 3.3.0.
  • #517 Added 'maxLength' metadata for Spark schema string fields.
  • #521 Fixed index generation and improved performance of variable block length files processing (record_format='VB').

• Thanks to the following people the project was made possible and for all the help along the way:

  • Andrew Baker, Francois Cillers, Adam Smyczek,
 Jan Scherbaum, Peter Moon, Clifford Lategan,
Rekha Gorantla, Mohit Suryavanshi, Niel Steyn
• Thanks to Tiago Requeijo, the author of the current ANTLR-based COBOL parser contributed to Cobrix.
• Thanks to the authors of the original COBOL parser. When we started the project we had zero knowledge of COBOL and this parser was a good starting point:
  • Ian De Beer, Rikus de Milander (https://github.com/zenaptix-lab/copybookStreams)

Companies, Names, Ids and values in all examples present in this project/repository are completely fictional and were generated randomly. Any resemblance to actual persons, companies or actual transactions is purely coincidental.

Take a look at other COBOL-related open source projects. If you think a project belongs in the list, please let us know, we will add it.

• RCOBOLDI - R COBOL DI (Data Integration) Package: An R package that facilitates the importation of COBOL CopyBook data directly into the R Project for Statistical Computing as properly structured data frames.