[GitHub] [hudi] the-other-tim-brown commented on a diff in pull request #6661: [HUDI-4853] Speeding up reading S3 files in S3EventsIncrSource

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

the-other-tim-brown commented on code in PR #6661:
URL: https://github.com/apache/hudi/pull/6661#discussion_r1008922752


##########
hudi-utilities/src/main/java/org/apache/hudi/utilities/sources/S3EventsHoodieIncrSource.java:
##########
@@ -213,15 +216,27 @@ public Pair<Option<Dataset<Row>>, String> fetchNextBatch(Option<String> lastCkpt
             }
           });
           return cloudFilesPerPartition.iterator();
-        }, Encoders.STRING()).collectAsList();
+        }, Encoders.STRING());
 
     Option<Dataset<Row>> dataset = Option.empty();
+    cloudFiles.cache();
     if (!cloudFiles.isEmpty()) {
-      DataFrameReader dataFrameReader = getDataFrameReader(fileFormat);
-      dataset = Option.of(dataFrameReader.load(cloudFiles.toArray(new String[0])));
+      JavaRDD<Dataset<Row>> datasetIterator = cloudFiles.javaRDD().flatMap(new FlatMapFunction<String, Dataset<Row>>() {
+        @Override
+        public Iterator<Dataset<Row>> call(String cloudFile) throws Exception {
+          return Collections.singletonList(getDataFromFile(cloudFile, fileFormat)).iterator();
+        }
+      });
+      Dataset<Row> finalDataset = datasetIterator.fold(sparkSession.emptyDataFrame(), (Function2<Dataset<Row>, Dataset<Row>, Dataset<Row>>) (v1, v2) -> v1.union(v2));
+      dataset = Option.of(finalDataset);
     }
-    LOG.debug("Extracted distinct files " + cloudFiles.size()
-        + " and some samples " + cloudFiles.stream().limit(10).collect(Collectors.toList()));
+    LOG.debug("Extracted distinct files " + cloudFiles.count()

Review Comment:
   I think this will always evaluate the count and collect the first ten elements to compute the value of the string before sending the value to the `LOG.debug` method. I think in this occasion we want to check if debug logs are enabled before executing this to avoid the extra overhead. 



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@hudi.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org