chenjunjiedada commented on code in PR #10200:
URL: https://github.com/apache/iceberg/pull/10200#discussion_r1576141345
##########
flink/v1.18/flink/src/main/java/org/apache/iceberg/flink/FlinkWriteConf.java:
##########
@@ -167,6 +168,18 @@ public DistributionMode distributionMode() {
return DistributionMode.fromName(modeName);
}
+ public DeleteGranularity deleteGranularity() {
+ String modeName =
+ confParser
+ .stringConf()
+ .option(FlinkWriteOptions.DELETE_GRANULARITY.key())
+ .flinkConfig(FlinkWriteOptions.DELETE_GRANULARITY)
+ .tableProperty(TableProperties.DELETE_GRANULARITY)
+ .defaultValue(TableProperties.DELETE_GRANULARITY_DEFAULT)
Review Comment:
Do we want to change to file granularity right now?
##########
data/src/test/java/org/apache/iceberg/io/TestTaskEqualityDeltaWriter.java:
##########
@@ -409,6 +421,55 @@ public void testUpsertDataWithFullRowSchema() throws
IOException {
.isEqualTo(ImmutableList.of(posRecord.copy("file_path",
dataFile.path(), "pos", 0L)));
}
+ @TestTemplate
+ public void testDeleteFileGranularity() throws IOException {
+ withGranularity(DeleteGranularity.FILE);
+ }
+
+ @TestTemplate
+ public void testDeletePartitionGranularity() throws IOException {
+ withGranularity(DeleteGranularity.PARTITION);
+ }
+
+ private void withGranularity(DeleteGranularity granularity) throws
IOException {
+ List<Integer> eqDeleteFieldIds = Lists.newArrayList(idFieldId,
dataFieldId);
+ Schema eqDeleteRowSchema = table.schema();
+
+ GenericTaskDeltaWriter deltaWriter =
+ createTaskWriter(eqDeleteFieldIds, eqDeleteRowSchema, granularity);
+
+ Map<Integer, Record> expected = Maps.newHashMapWithExpectedSize(2000);
+ // Create enough records, so we have multiple files
+ for (int i = 0; i < 2000; ++i) {
+ Record record = createRecord(i, "aaa" + i);
+ deltaWriter.write(record);
+ if (i % 5 == 10) {
+ deltaWriter.delete(record);
+ } else {
+ expected.put(i, record);
+ }
+ }
+
+ // Add some deletes in the end
+ for (int i = 0; i < 199; ++i) {
+ int id = i * 10 + 1;
+ Record record = createRecord(id, "aaa" + id);
+ deltaWriter.delete(record);
+ expected.remove(id);
+ }
+
+ WriteResult result = deltaWriter.complete();
+ assertThat(result.dataFiles()).as("Should have 2 data files.").hasSize(2);
Review Comment:
Just recall this.
##########
flink/v1.18/flink/src/main/java/org/apache/iceberg/flink/sink/FlinkSink.java:
##########
@@ -242,6 +243,18 @@ public Builder distributionMode(DistributionMode mode) {
return this;
}
+ /**
+ * Configure the write {@link DeleteGranularity} that the flink sink will
use. Currently, flink
+ * support {@link DeleteGranularity#PARTITION} and {@link
DeleteGranularity#FILE}.
Review Comment:
How about documenting the default granularity that Flink sink uses?
--
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: issues-unsubscr...@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org
For additional commands, e-mail: issues-h...@iceberg.apache.org
