Writing of FMT_MAINFRAME adds 2 bytes (0x0d0x25) to every line

[dirk jaeckel]

Hello Bruce,
first at all: thank you for the great work you are doing here!

I try to write a file with the following code(dont wonder about the fully qualified names, its nessecary in talend):

net.sf.JRecord.Details.LayoutDetail schema = net.sf.JRecord.utilityClasses.SchemaLoader.loadSchema(     "V:/paketierung/A244_WPL/01_Dokumentation/02_Datenmodell/Copybooks/cobol/ACTIVITY-TAB",
    net.sf.JRecord.External.CopybookLoader.SPLIT_NONE, 
    "cp1141", 
    net.sf.JRecord.Numeric.ICopybookDialects.FMT_MAINFRAME).asLayoutDetail();
    //
    int recordIdx = 0;  // since there is only one record type, the record index must be zero
                         // If there where more than one record, you could use:
                         //    schema.getRecordIndex("COMPANY-RECORD");

        net.sf.JRecord.IO.AbstractLineWriter  writer = net.sf.JRecord.IO.LineIOProvider
                        .getInstance().getLineWriter(schema);

        writer.open(context.file);
        net.sf.JRecord.Details.AbstractLine line = new net.sf.JRecord.Details.Line(schema);

        String[] fieldNames = line.getLayout().getFieldDescriptions(0,0);
        System.err.println(java.util.Arrays.deepToString(fieldNames));

        // Retrieve the Record-Definition
        net.sf.JRecord.Details.RecordDetail  record = schema.getRecord(recordIdx);
        System.err.println("Schemainfos:");
        System.err.println("schema.getFileStructure "+ schema.getFileStructure());
        System.err.println("schema.getFontName "+schema.getFontName());
        System.err.println("schema.isBinary "+ schema.isBinary());

for creating the file
Output:
-------------------------------------------------------
Warning there was no "cb2xml.properties" file,
using the default Cobol columns of 6 to 72
-------------------------------------------------------
*** using start column = 6, end column = 72
Schemainfos:
schema.getFileStructure 1
schema.getFontName cp1141
schema.isBinary false

then i fill 61 lines and write them :

// this codepice is triggerd 61 times
if (record.getFieldCount()==row2.alles.getColumnCount() -1) {
    // row2.alles is a dynamic structure with an additional linenumber in pos 0
    for (int idx = 1, count = row2.alles.getColumnCount(); idx < count; ++idx) {
            line.getFieldValue(record.getField(idx-1)).set(row2.alles.getColumnValue(idx));
    }
    System.out.println("|"+line.getFullLine()+"|");
    writer.write(line);
}  else {
    throw new Exception("Spaltendefinition zwischen Tabelle und Copybook unterschiedlich");// diffrence in definitions detected

}

The system out println of getFullLine gives me 100% what i expect !
After closing the file

// end of the component, outside/closing the loop
    writer.close();  

i load the file with Recordeditor and the same Copybook, i find between the records: 0d25

What am i doing wrong ?

Regards
Dirk

[dirk jaeckel]

Hi again, sorry for posting in the wrong section, anyway, the library version i use is:
JRecord_Version_0.81.4
which includes:
Manifest-Version: 1.0
Ant-Version: Apache Ant 1.8.4
Created-By: 1.7.0_79-b15 (Oracle Corporation)
Main-Class: net.sf.cb2xml.Cb2Xml
Implementation-Version: 0.95.8
Built-Date: 2016-11-05 16:43:51

[dirk jaeckel]

I found the error: i used a line writer:

net.sf.JRecord.IO.LineIOProvide.getInstance().getLineWriter(schema);

should have done this:

   ISchemaIOBuilder cobolBuilder = JRecordInterface1.COBOL
           .newIOBuilder(context.copybook)
            .setFileOrganization(net.sf.JRecord.Common.Constants.IO_FIXED_LENGTH)
            .setDialect(net.sf.JRecord.Numeric.ICopybookDialects.FMT_MAINFRAME);

    AbstractLineWriter writer =  cobolBuilder.newWriter(context.file);
    LayoutDetail  schema = cobolBuilder.getLayout();

a little confusing the start with only javadoc in hand, but really great work from your side :)

[Bruce Martin]

Good point, I will update JavaDoc to reference JRecordInterface1.

In the JRecord source directory there is JRecord_IO_Builder_Examples which holds examples of using JRecord.

Also in the RecordEditor, the Generate >>> Java JRecord Code for Cobol option will generate basic example JRecord code from a Cobol copybook (and optionally data file). See file:///F:/Work/JRecord/Versions/Version_0.81/Docs/RecordEditorGenerate.htm#HDRJRECGEN for details.

If you are editing a CSV file in the RecordEditor, you can use the Generate option to generate JRecord code to read / write the CSV file. I plan to add these options to ReCsvEditor as well.