From 10af25f58982fc7284f05cb80b12126cf3ee95b7 Mon Sep 17 00:00:00 2001
From: Dennis Schafroth <dennis@indexdata.com>
Date: Wed, 4 Jan 2012 20:14:47 +0100
Subject: [PATCH] Possible to skip bad indicators. Replace control characters
 with space.

---
 src/org/marc4j/MarcStreamReader.java |   34 +++++++++++++++++++++++++++++++---
 1 file changed, 31 insertions(+), 3 deletions(-)

diff --git a/src/org/marc4j/MarcStreamReader.java b/src/org/marc4j/MarcStreamReader.java
index 473450a..9b6f637 100644
--- a/src/org/marc4j/MarcStreamReader.java
+++ b/src/org/marc4j/MarcStreamReader.java
@@ -83,6 +83,8 @@ public class MarcStreamReader implements MarcReader {
        
     private CharConverter converterAnsel = null;
 
+    private boolean setBadIndicators = true;
+
     /**
      * Constructs an instance with the specified input stream.
      */
@@ -234,7 +236,10 @@ public class MarcStreamReader implements MarcReader {
                     inputrec.readFully(byteArray);
     
                     try {
-                        record.addVariableField(parseDataField(tags[i], byteArray));
+                      DataField dataField = parseDataField(tags[i], byteArray);
+                      // dataField could be null if bad indicators
+                      if (dataField != null) 
+                	record.addVariableField(dataField);
                     } catch (IOException e) {
                         throw new MarcException(
                                 "error parsing data field for tag: " + tags[i]
@@ -263,9 +268,16 @@ public class MarcStreamReader implements MarcReader {
 
         DataField dataField = factory.newDataField();
         dataField.setTag(tag);
-        dataField.setIndicator1(ind1);
-        dataField.setIndicator2(ind2);
 
+        boolean badIndicatorFound = false;
+        if (setBadIndicators || ind1 >= ' ' )
+          dataField.setIndicator1(ind1);
+        else
+          badIndicatorFound = true;
+        if (setBadIndicators || ind2 >= ' ')
+          dataField.setIndicator2(ind2);
+        else
+          badIndicatorFound = true;
         int code;
         int size;
         int readByte;
@@ -294,6 +306,9 @@ public class MarcStreamReader implements MarcReader {
                 break;
             }
         }
+        /* Bad Indicators was found, so dropping field */
+        if (badIndicatorFound)
+          return null;
         return dataField;
     }
     
@@ -406,7 +421,12 @@ public class MarcStreamReader implements MarcReader {
         else if (encoding.equals("MARC-8") || encoding.equals("MARC8"))
         {
             if (converterAnsel == null) converterAnsel = new AnselToUnicode();
+
+            for (int index = 0; index < bytes.length; index++)
+              if (bytes[index] < 32)
+        	bytes[index] = ' ';
             dataElement = converterAnsel.convert(bytes);
+            //dataElement = dataElement.replaceAll("\0", " ");
         }
         else if (encoding.equals("ISO-8859-1") || encoding.equals("ISO8859_1"))
         {
@@ -419,5 +439,13 @@ public class MarcStreamReader implements MarcReader {
         }
         return dataElement;
     }
+
+    public boolean isBadIndicators() {
+      return setBadIndicators;
+    }
+
+    public void setBadIndicators(boolean trueFalse) {
+      this.setBadIndicators = trueFalse;
+    }
     
 }
\ No newline at end of file
-- 
1.7.10.4