Improved record id extraction from field 003@$0.

The new code handles record ids properly if they are at the end of the
record and not terminated by a field or subfield delimiter.

Author: cboehme

src/main/java/org/culturegraph/mf/stream/converter/bib/PicaDecoder.java

@@ -158,6 +158,20 @@ private boolean recordIsEmpty() {
 		return true;
 	}
 
+	/**
+	 * Searches the record for the sequence specified in {@code ID_FIELD}
+	 * and returns all characters following this sequence until the next
+	 * {@link PicaConstants.FIELD_DELIMITER},
+	 * {@link PicaConstants.SUBFIELD_DELIMITER} or the end of the record
+	 * is reached. Only the first occurrence of the sequence is processed,
+	 * later occurrences are ignored.
+	 * 
+	 * If the sequence is not found in the string or if it is not followed
+	 * by any characters then {@code null} is returned.
+	 * 
+	 * @return value of subfield 003@$0 or null if the
+	 *         field is not found or is empty.
+	 */
 	private String extractRecordId() {
 		idBuilder.setLength(0);
 
@@ -166,29 +180,31 @@ private String extractRecordId() {
 		for (int i = 0; i < recordLen; ++i) {
 			if (buffer[i] == PicaConstants.FIELD_DELIMITER) {
 				if (idBuilder.length() > 0) {
-					return idBuilder.toString();
+					break;
 				}
 				fieldPos = 0;
 				skip = false;
-				continue;
-			}
-			if (!skip) {
-				if (fieldPos < ID_FIELD.length) {
-					if (buffer[i] == ID_FIELD[fieldPos]) {
-						fieldPos += 1;
-					} else {
-						skip = true;
-					}
-				} else {
-					if (buffer[i] == PicaConstants.SUBFIELD_DELIMITER) {
-						skip = true;
+			} else {
+				if (!skip) {
+					if (fieldPos < ID_FIELD.length) {
+						if (buffer[i] == ID_FIELD[fieldPos]) {
+							fieldPos += 1;
+						} else {
+							skip = true;
+						}
 					} else {
+						if (buffer[i] == PicaConstants.SUBFIELD_DELIMITER) {
+							break;
+						}
 						idBuilder.append(buffer[i]);
 					}
 				}
 			}
 		}
 
+		if (idBuilder.length() > 0) {
+			return idBuilder.toString();
+		}
 		return null;
 	}
 

src/test/java/org/culturegraph/mf/stream/converter/bib/PicaDecoderTest.java

@@ -41,8 +41,9 @@ public final class PicaDecoderTest {
 	private static final String SUBFIELD = "\u001f";
 	private static final String FIELD_END = "\u001e";
 
+	private static final String FIELD_003AT0_START = "003@ " + SUBFIELD + "0";
 	private static final String FIELD_001AT = "001@ " + SUBFIELD + "0test" + FIELD_END;
-	private static final String FIELD_003AT = "003@ " + SUBFIELD + "0" + RECORD_ID + FIELD_END;
+	private static final String FIELD_003AT = FIELD_003AT0_START + RECORD_ID + FIELD_END;
 	private static final String FIELD_021A = "021A " + SUBFIELD + "a" + COMPOSED_UTF8 + FIELD_END;
 	private static final String FIELD_028A_START = "028A ";
 	private static final String SUBFIELD_A = SUBFIELD + "aEco";
@@ -237,6 +238,20 @@ public void testShouldNotSkipEmptyFieldsWithOnlyEmptySubfieldsIfConfigured() {
 		verify028AEnd(ordered);
 		ordered.verify(receiver).endRecord();
 	}
+	
+	@Test
+	public void testShouldProcessRecordIdAtRecordEnd() {
+		picaDecoder.setFixUnexpectedEOR(true);
+
+		picaDecoder.process(
+				FIELD_003AT0_START +
+				RECORD_ID);
+
+		final InOrder ordered = inOrder(receiver);
+		ordered.verify(receiver).startRecord(RECORD_ID);
+		verify003At(ordered);
+		ordered.verify(receiver).endRecord();
+	}
 
 	@Test(expected=MissingIdException.class)
 	public void testShouldFailIfIdIsMissingByDefault() {