[GitHub] markap14 commented on a change in pull request #3282: NIFI-5983: handling parse problems in recordReader implementations

[GitBox <gi...@apache.org>]

markap14 commented on a change in pull request #3282: NIFI-5983: handling parse problems in recordReader implementations
URL: https://github.com/apache/nifi/pull/3282#discussion_r253565293
 
 

 ##########
 File path: nifi-nar-bundles/nifi-standard-services/nifi-record-serialization-services-bundle/nifi-record-serialization-services/src/main/java/org/apache/nifi/avro/AvroRecordReader.java
 ##########
 @@ -33,14 +35,21 @@
 
     @Override
     public Record nextRecord(final boolean coerceTypes, final boolean dropUnknownFields) throws IOException, MalformedRecordException {
-        GenericRecord record = nextAvroRecord();
-        if (record == null) {
-            return null;
+        try {
+            GenericRecord record = nextAvroRecord();
+            if (record == null) {
+                return null;
+            }
+
+            final RecordSchema schema = getSchema();
+            final Map<String, Object> values = AvroTypeUtil.convertAvroRecordToMap(record, schema);
+            return new MapRecord(schema, values);
+        } catch (IOException e) {
+            throw e;
+        } catch (MalformedRecordException e) {
+            throw e;
+        } catch (Throwable e) {
+            throw new MalformedRecordException(String.format("Error while getting next record. Root cause: %s", Throwables.getRootCause(e).getMessage()), e);
 
 Review comment:
   We should avoid `String.format` and instead just concatenate the two Strings. `String.format` is quite expensive, parsing a Regular Expression behind the scenes, then evaluating it against the String and performing substitution. If formatting of numbers of different precisions, etc. is required then it makes sense to do so but in a case where it's just concatenating Strings I would generally steer clear of it.
   
   I would also avoid using `Throwables.getRootCause(e).getMessage()` and instead just use `Throwables.getRootCause(e)`. Often times, especially for RuntimeException's, such as NullPointerException, NumberFormatException, etc, there is no message, or the message only makes sense in the context of the Exception. For example, if a value is expected to be a number but is the value "hello", we could get a `NumberFormatException` with a message of `hello`. In this case, the Exception would read: `Error while getting next record. Root cause: hello`. Instead, keeping the `toString()` of the Exception would yield `Error while getting next record. Root cause: NumberFormatException: hello`

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services