Yizhou-Yang commented on code in PR #7418:
URL: https://github.com/apache/inlong/pull/7418#discussion_r1116659813
##########
inlong-sort/sort-connectors/iceberg/src/main/java/org/apache/inlong/sort/iceberg/sink/multiple/IcebergMultipleStreamWriter.java:
##########
@@ -201,7 +239,56 @@ public void processElement(RecordWithSchema
recordWithSchema) throws Exception {
if (multipleWriters.get(tableId) != null) {
for (RowData data : recordWithSchema.getData()) {
- multipleWriters.get(tableId).processElement(data);
+ try {
+ multipleWriters.get(tableId).processElement(data);
+ } catch (Exception e) {
+ LOG.error(String.format("write error, raw data: %s",
data), e);
+ if (!dirtyOptions.ignoreDirty()) {
+ throw e;
+ }
+ if (dirtySink != null) {
+ DirtyData.Builder<Object> builder =
DirtyData.builder();
+ try {
+ String dataBaseName =
tableId.namespace().toString();
+ String tableName = tableId.name();
+ String dirtyLabel =
DirtySinkHelper.regexReplace(dirtyOptions.getLabels(),
+ DirtyType.BATCH_LOAD_ERROR, null,
+ dataBaseName, tableName, null);
+ String dirtyLogTag =
+
DirtySinkHelper.regexReplace(dirtyOptions.getLogTag(),
+ DirtyType.BATCH_LOAD_ERROR, null,
+ dataBaseName, tableName, null);
+ String dirtyIdentifier =
+
DirtySinkHelper.regexReplace(dirtyOptions.getIdentifier(),
+ DirtyType.BATCH_LOAD_ERROR, null,
+ dataBaseName, tableName, null);
+ builder.setData(data)
+ .setLabels(dirtyLabel)
+ .setLogTag(dirtyLogTag)
+ .setIdentifier(dirtyIdentifier)
+
.setRowType(multipleWriters.get(tableId).getFlinkRowType())
+ .setDirtyMessage(e.getMessage());
+ dirtySink.invoke(builder.build());
+ if (sinkMetricData != null) {
+ long size =
data.toString().getBytes(StandardCharsets.UTF_8).length;
+
sinkMetricData.outputDirtyMetricsWithEstimate(dataBaseName,
+ tableName, 1, size);
+ }
+ } catch (Exception ex) {
+ if (!dirtyOptions.ignoreSideOutputErrors()) {
+ throw new RuntimeException(ex);
+ }
+ LOG.warn("Dirty sink failed", ex);
+ }
+ }
+ }
+
+ if (sinkMetricData != null) {
+ String dataBaseName = tableId.namespace().toString();
+ String tableName = tableId.name();
+ long size =
data.toString().getBytes(StandardCharsets.UTF_8).length;
+ sinkMetricData.outputMetrics(dataBaseName, tableName, 1,
size);
Review Comment:
either add dirty metric or clean metric, your current implementation might
add both metrics at once.
consider adding a return clause at the end of line 275 or something, so that
only one of sinkMetricData.outputMetrics and
sinkMetricData.outputDirtyMetricsWithEstimate is called
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
