Problema with record ID

[svrolanski]

If I have a fixed length records like this

Xmessage
XDother message
XDanother message

Where X, XD and XR are the Ids, BeanIO cannot identify the record type correctly.

[nicoschl]

@svrolanski

Please provide your mapping.xml file or the equivalent code with the annotations

[svrolanski]

class RecordX {
	public static final String ID_REG = "X";
	@Field(at = 0, length = 1, rid = true, literal = ID_REG)
	private String idReg;

class RecordXD {
	public static final String ID_REG = "XD";
	@Field(at = 0, length = 2, rid = true, literal = ID_REG)
	private String idReg;

class RecordXR {
	public static final String ID_REG = "XR";
	@Field(at = 0, length = 2, rid = true, literal = ID_REG)
	private String idReg;

[nicoschl]

Your classes doesn't specify any min/max occurrences. How are your data then structured?
Is there only a single RecordX message and then any number of RecordXD and RecordXR records?

How do you define your StreamBuilder?

The above work for me with a simple change to RecordX:

@Record(minOccurs = 1, maxOccurs = 1)
public class RecordX {

Is the test data correct? Should the last message not start with XR? More example data could also be useful if the above does not answer/solve your problem.

[svrolanski]

There are any number (0..unbonded) of X/XD/XR records in any order. I just "new StreamBuilder()" and call addRecord(RecordX.class), ... for each class. The order I call addRecord() does not seem to matter also. I workaround the problem by creating two streamBuilder and then creating two unmarshallesr. I can check for the record type myself and use the correct unmarshaller:

public <T extends Record> T toRecord(String line) {
	if (line.startsWith(RecordXD.ID_REG) || line.startsWith(RecordXR.ID_REG))
		// has only RecordXD RecordXR
		return (T) unmarshallerXdXr.unmarshal(line);

	// has only RecordX
	return (T) unmarshaller.unmarshal(line);
}

Anyway, this definitely looks like a bug to me. Since sometimes you cannot control how the input file is created, you need to predict such situations where IDs might start with the same string.