Merge pull request #212 from metamx/realtime-npe

Rejigger exception handling in standalone realtime

indexing-common/src/main/java/com/metamx/druid/indexer/data/InputRowParser.java

 package com.metamx.druid.indexer.data;
 
+import com.metamx.common.exception.FormattedException;
 import com.metamx.druid.input.InputRow;
 
 public interface InputRowParser<T>
 {
-  public InputRow parse(T input);
+  public InputRow parse(T input) throws FormattedException;
   public void addDimensionExclusion(String dimension);
 }

indexing-common/src/main/java/com/metamx/druid/indexer/data/MapInputRowParser.java

@@ -4,6 +4,7 @@ import com.fasterxml.jackson.annotation.JsonCreator;
 import com.fasterxml.jackson.annotation.JsonProperty;
 import com.google.common.collect.Lists;
 import com.google.common.collect.Sets;
+import com.metamx.common.exception.FormattedException;
 import com.metamx.druid.input.InputRow;
 import com.metamx.druid.input.MapBasedInputRow;
 import org.joda.time.DateTime;
@@ -37,13 +38,15 @@ public class MapInputRowParser implements InputRowParser<Map<String, Object>>
   }
 
   @Override
-  public InputRow parse(Map<String, Object> theMap)
+  public InputRow parse(Map<String, Object> theMap) throws FormattedException
   {
     final List<String> dimensions = dataSpec.hasCustomDimensions()
                                     ? dataSpec.getDimensions()
                                     : Lists.newArrayList(Sets.difference(theMap.keySet(), dimensionExclusions));
 
-    final DateTime timestamp = timestampSpec.extractTimestamp(theMap);
+    final DateTime timestamp;
+    try {
+      timestamp = timestampSpec.extractTimestamp(theMap);
       if (timestamp == null) {
         final String input = theMap.toString();
         throw new NullPointerException(
@@ -53,6 +56,13 @@ public class MapInputRowParser implements InputRowParser<Map<String, Object>>
             )
         );
       }
+    }
+    catch (Exception e) {
+      throw new FormattedException.Builder()
+          .withErrorCode(FormattedException.ErrorCode.UNPARSABLE_TIMESTAMP)
+          .withMessage(e.toString())
+          .build();
+    }
 
     return new MapBasedInputRow(timestamp.getMillis(), dimensions, theMap);
   }

realtime/src/main/java/com/metamx/druid/realtime/RealtimeManager.java

@@ -164,6 +164,17 @@ public class RealtimeManager implements QuerySegmentWalker
           final InputRow inputRow;
           try {
             inputRow = firehose.nextRow();
+          }
+          catch (FormattedException e) {
+            log.info(e, "unparseable line: %s", e.getDetails());
+            metrics.incrementUnparseable();
+            continue;
+          }
+          catch (Exception e) {
+            log.info(e, "thrown away line due to exception");
+            metrics.incrementThrownAway();
+            continue;
+          }
 
           final Sink sink = plumber.getSink(inputRow.getTimestampFromEpoch());
           if (sink == null) {
@@ -185,12 +196,6 @@ public class RealtimeManager implements QuerySegmentWalker
             nextFlush = new DateTime().plus(intermediatePersistPeriod).getMillis();
           }
         }
-          catch (FormattedException e) {
-            log.info(e, "unparseable line: %s", e.getDetails());
-            metrics.incrementUnparseable();
-            continue;
-          }
-        }
       } catch (RuntimeException e) {
         log.makeAlert(e, "RuntimeException aborted realtime processing[%s]", fireDepartment.getSchema().getDataSource())
            .emit();