Merge pull request #441 from metafacture/440-doNotTrimControlField

Do not trim control field values in MARCXML handler.
metafacture · Feb 17, 2022 · 7799c1e · 7799c1e
2 parents 89793c7 + df6a862
commit 7799c1e
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 1 deletion.
diff --git a/metafacture-biblio/src/main/java/org/metafacture/biblio/marc21/MarcXmlHandler.java b/metafacture-biblio/src/main/java/org/metafacture/biblio/marc21/MarcXmlHandler.java
@@ -126,7 +126,7 @@ else if (DATAFIELD.equals(localName)) {
             getReceiver().endEntity();
         }
         else if (CONTROLFIELD.equals(localName)) {
-            getReceiver().literal(currentTag, builder.toString().trim());
+            getReceiver().literal(currentTag, builder.toString());
         }
         else if (RECORD.equals(localName) && checkNamespace(uri)) {
             getReceiver().endRecord();

diff --git a/metafacture-biblio/src/test/java/org/metafacture/biblio/marc21/MarcXmlHandlerTest.java b/metafacture-biblio/src/test/java/org/metafacture/biblio/marc21/MarcXmlHandlerTest.java
@@ -77,6 +77,20 @@ public void shouldFindTagAttributeAtSecondPositionInControlFieldElement()
         verify(receiver).literal("001", fieldValue);
     }
 
+    @Test
+    public void issue440_shouldNotRemoveWhitespaceFromControlFields() throws SAXException {
+        final AttributesImpl attributes = new AttributesImpl();
+        attributes.addAttribute(NAMESPACE, "tag", "tag", "CDATA", "008");
+
+        final String fieldValue = "      t20202020au |||||||||||| ||||ger d";
+
+        marcXmlHandler.startElement(NAMESPACE, CONTROLFIELD, "", attributes);
+        marcXmlHandler.characters(fieldValue.toCharArray(), 0, fieldValue.length());
+        marcXmlHandler.endElement(NAMESPACE, CONTROLFIELD, "");
+
+        verify(receiver).literal("008", fieldValue);
+    }
+
     @Test
     public void issue233ShouldNotRemoveWhitespaceFromLeader()
             throws SAXException {