This release contains extra functions for the CodeGen utility, Csv parsing changes and fix for duplicate field processing.
These are to reduce the amount of duplicate code int JRecord/RecordEditor.
The ReadMe_NewUsers.md has an expanded version of this section. The first thing new users need to decide is
- Are you going to use the Java interface ???, then see Java Interface section
- Are you going to use the existing Cobol2Csv / Cobol2Xml interfaces. These programs will convert Cobol-Data-Files to/from Csv/Xml files.
- Are you going to write your own generic interface ???. If so look at the CobolSchemaReader class and see how it is used in Cobol2Xml. Also If copying files, look at Copy class.
JRecord is available from Sourceforge:
- JRecord on Sourceforge
- JRecord Download
- JRecord Questions on stackoverflow
- JRecord Forums on Sourceforge
- JRecord Tickets on Sourceforge
JRecord provides
- IO routines for Length Based (Lines are either fixed length or a length field at the start of the record) files. Standard Text files are also supported
- IO routines for Cobol Data files (both Text and Binary).
- IO routines for Flat Files (both Fixed Field width and CSV files).
- Cobol2Csv, Csv2Cobol and Cobol2Cobol utilities
- Same IO routines for all File Types (XML (via StAX parser), CSV and Fixed Field Width)
For Documentation see JRecord Documentation
The easiest way to access JRecord is via IOBuilders (using the JRecordInterface1 class):
ICobolIOBuilder ioBldr
= JRecordInterface1.COBOL
.newIOBuilder("CobolCopybook.cbl")
.setSplitCopybook(CopybookLoader.SPLIT_01_LEVEL)
.setDialect(ICopybookDialects.FMT_FUJITSU);
AbstractLineReader reader = ioBldr.newReader("input File Name");
AbstractLine l;
while ((l = reader.read()) != null) {
...
}
reader.close();There is a description of using the IOBuilder interface in the JRecord Manual. You should also investigate the Code Generation options.
To help you get started with JRecord, there are 2 Code Generate options available:
-
The JRecord CodeGen sub-project will generate a skelton Java~JRecord Read/write programs from a Cobol Copybook.
-
JRecord CodeGen is now part of the RecordEditor see RecordEditor Generate option for details. Also see
Code to read a Cobol file (Generated by CodeGen utility):
ICobolIOBuilder iob = JRecordInterface1.COBOL
.newIOBuilder(copybookName)
.setFont("cp037")
.setFileOrganization(Constants.IO_FIXED_LENGTH)
.setSplitCopybook(CopybookLoader.SPLIT_NONE)
;
FieldNamesDtar021.RecordDtar021 rDtar021 = FieldNamesDtar021.RECORD_DTAR021;
AbstractLineReader reader = iob.newReader(dataFile);
while ((line = reader.read()) != null) {
lineNum += 1;
System.out.println(
line.getFieldValue(rDtar021.keycodeNo).asString()
+ " " + line.getFieldValue(rDtar021.storeNo).asString()
+ " " + line.getFieldValue(rDtar021.qtySold).asString()
+ " " + line.getFieldValue(rDtar021.salePrice).asString()
);
}
reader.close();Cobol Field Name class (Generated by CodeGen utility):
public class FieldNamesDtar021 {
public static final RecordDtar021 RECORD_DTAR021 = new RecordDtar021();
public static class RecordDtar021 {
public final String keycodeNo = "KEYCODE-NO";
public final String storeNo = "STORE-NO";
public final String theDate = "THE-DATE";
public final String deptNo = "DEPT-NO";
public final String qtySold = "QTY-SOLD";
public final String salePrice = "SALE-PRICE";
}
} FieldNamesDtar022.RecordDtar022 rDtar022 = FieldNamesDtar022.RECORD_DTAR022;
ICobolIOBuilder iobWrite = JRecordInterface1.COBOL
.newIOBuilder(outCopybookName)
.setFont("cp037")
.setFileOrganization(Constants.IO_FIXED_LENGTH)
.setSplitCopybook(CopybookLoader.SPLIT_NONE)
;
AbstractLineWriter writer = iobWrite.newWriter(outputFileName);
AbstractLine dtar022Line = iobWrite.newLine();
dtar022Line.getFieldValue(rDtar022.keycodeNo).set("223");
dtar022Line.getFieldValue(rDtar022.theDate) .set(22);
dtar022Line.getFieldValue(rDtar022.deptNo) .set(22);
dtar022Line.getFieldValue(rDtar022.qtySold) .set(5);
dtar022Line.getFieldValue(rDtar022.salePrice).set("123.45");
writer.write(dtar022Line);
...
writer.close();If working with Hadoop have a look at:
Previous versions of JRecord came with utilities (editors etc), this version does not with old utilities (it does have Csv2Cobol and Cobol2Csv utilities).
I would suggest you:
- Download and use the RecordEditor to view / edit your Cobol files. You can also use the RecordEditor to generate JRecord code.
You can use the RecordEditor to view / try out the JRecord-options before writing your code. In the future, I hope to release a Cobol-Version of the RecordEditor that uses Cobol-Copybooks and Xml-File descriptions. This will replace the old JRecord-Utilities. This is unlikely to be released before the end of 2015 though.
- Change to use cb2xml 1.00. Cb2xml now converts Cobol directly into Java classes which are used by JRecord (instead of going via Xml).
- JRecord now stores both the traditional Flat field list + Cobol Item Tree. This provides greater access to the original Cobol data.
- Improved performance when accessing Comp, Comp-3 and Mainframe Zoned Decimal. For smaller numerics, conversion is done using long variables instead of Objects. This improves performance by 70 / 80 %
- New Cobol Copybook Split option.
- Updates to CodeGen - JRecord Code Generator.
- Fix to support large copybooks
-
Minor improvements in handling duplicate field names
-
Csv changes to Bring JRecord into line with the RecordEditor, in particular
- Csv Quotes can now be expressed in hex format x'FF' etc.
- JRecord has both Chartacter based Csv Parser and a Byte based Csv parser. The Byte based Csv was very basic but it has been upgraded to be the same as the Chartacter based Csv Parser.
-
Extra Functionality For use by CodeGen generated code.
-
Two new Code Templates added to the CodeGen utility.
Changes for complex Occurs Depending
- Reduction in Memory usage in CobolToXml, CobolToJson when using Occurs depending.
- Fixes for Complex Occurs-Depend for CobolToXml, CobolToJson
- Added Pretty-print and asIoBuilder options to CobolToJson builder
- Add pom files to CobolToJson, Fix pom files for CobolToXml
- Csv-field-delimiters can be specified in \u0001 format (unicode hex format
- Some Csv fix's for specifying field-delimiters in the x'01' type format.