lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

[HUDI-2656] Generalize HoodieIndex for flexible record data type (#3893)

Browse Source 
Co-authored-by: Raymond Xu <2701446+xushiyan@users.noreply.github.com>
This commit is contained in:
Y Ethan Guo
2022-02-03 20:24:04 -08:00
committed by GitHub
parent 69dfcda116
commit b8601a9f58
105 changed files with 564 additions and 504 deletions
Download Patch File Download Diff File Expand all files Collapse all files

39
hudi-client/hudi-flink-client/src/test/java/org/apache/hudi/index/bloom/TestFlinkHoodieBloomIndex.java
View File

@@ -23,6 +23,7 @@ import org.apache.hudi.common.bloom.BloomFilterFactory;
import org.apache.hudi.common.bloom.BloomFilterTypeCode;
import org.apache.hudi.common.data.HoodieList;
import org.apache.hudi.common.data.HoodieMapPair;
import org.apache.hudi.common.model.HoodieAvroRecord;
import org.apache.hudi.common.model.HoodieKey;
import org.apache.hudi.common.model.HoodieRecord;
import org.apache.hudi.common.table.HoodieTableMetaClient;
@@ -115,19 +116,19 @@ public class TestFlinkHoodieBloomIndex extends HoodieFlinkClientTestHarness {
RawTripTestPayload rowChange1 =
new RawTripTestPayload("{\"_row_key\":\"000\",\"time\":\"2016-01-31T03:16:41.415Z\",\"number\":12}");
HoodieRecord record1 =
new HoodieRecord(new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath()), rowChange1);
new HoodieAvroRecord(new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath()), rowChange1);
RawTripTestPayload rowChange2 =
new RawTripTestPayload("{\"_row_key\":\"001\",\"time\":\"2016-01-31T03:16:41.415Z\",\"number\":12}");
HoodieRecord record2 =
new HoodieRecord(new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath()), rowChange2);
new HoodieAvroRecord(new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath()), rowChange2);
RawTripTestPayload rowChange3 =
new RawTripTestPayload("{\"_row_key\":\"002\",\"time\":\"2016-01-31T03:16:41.415Z\",\"number\":12}");
HoodieRecord record3 =
new HoodieRecord(new HoodieKey(rowChange3.getRowKey(), rowChange3.getPartitionPath()), rowChange3);
new HoodieAvroRecord(new HoodieKey(rowChange3.getRowKey(), rowChange3.getPartitionPath()), rowChange3);
RawTripTestPayload rowChange4 =
new RawTripTestPayload("{\"_row_key\":\"003\",\"time\":\"2016-01-31T03:16:41.415Z\",\"number\":12}");
HoodieRecord record4 =
new HoodieRecord(new HoodieKey(rowChange4.getRowKey(), rowChange4.getPartitionPath()), rowChange4);
new HoodieAvroRecord(new HoodieKey(rowChange4.getRowKey(), rowChange4.getPartitionPath()), rowChange4);
List<String> partitions = asList("2016/01/21", "2016/04/01", "2015/03/12");
List<Pair<String, BloomIndexFileInfo>> filesList = index.loadColumnRangesFromFiles(partitions, context, hoodieTable);
@@ -212,16 +213,16 @@ public class TestFlinkHoodieBloomIndex extends HoodieFlinkClientTestHarness {
+ "\"time\":\"2016-01-31T03:16:41.415Z\",\"number\":32}";
RawTripTestPayload rowChange1 = new RawTripTestPayload(recordStr1);
HoodieRecord record1 =
new HoodieRecord(new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath()), rowChange1);
new HoodieAvroRecord(new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath()), rowChange1);
RawTripTestPayload rowChange2 = new RawTripTestPayload(recordStr2);
HoodieRecord record2 =
new HoodieRecord(new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath()), rowChange2);
new HoodieAvroRecord(new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath()), rowChange2);
RawTripTestPayload rowChange3 = new RawTripTestPayload(recordStr3);
HoodieRecord record3 =
new HoodieRecord(new HoodieKey(rowChange3.getRowKey(), rowChange3.getPartitionPath()), rowChange3);
new HoodieAvroRecord(new HoodieKey(rowChange3.getRowKey(), rowChange3.getPartitionPath()), rowChange3);
RawTripTestPayload rowChange4 = new RawTripTestPayload(recordStr4);
HoodieRecord record4 =
new HoodieRecord(new HoodieKey(rowChange4.getRowKey(), rowChange4.getPartitionPath()), rowChange4);
new HoodieAvroRecord(new HoodieKey(rowChange4.getRowKey(), rowChange4.getPartitionPath()), rowChange4);
// We write record1, record2 to a base file, but the bloom filter contains (record1,
// record2, record3).
@@ -286,16 +287,16 @@ public class TestFlinkHoodieBloomIndex extends HoodieFlinkClientTestHarness {
String recordStr4 = "{\"_row_key\":\"" + rowKey1 + "\",\"time\":\"2015-01-31T03:16:41.415Z\",\"number\":32}";
RawTripTestPayload rowChange1 = new RawTripTestPayload(recordStr1);
HoodieRecord record1 =
new HoodieRecord(new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath()), rowChange1);
new HoodieAvroRecord(new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath()), rowChange1);
RawTripTestPayload rowChange2 = new RawTripTestPayload(recordStr2);
HoodieRecord record2 =
new HoodieRecord(new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath()), rowChange2);
new HoodieAvroRecord(new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath()), rowChange2);
RawTripTestPayload rowChange3 = new RawTripTestPayload(recordStr3);
HoodieRecord record3 =
new HoodieRecord(new HoodieKey(rowChange3.getRowKey(), rowChange3.getPartitionPath()), rowChange3);
new HoodieAvroRecord(new HoodieKey(rowChange3.getRowKey(), rowChange3.getPartitionPath()), rowChange3);
RawTripTestPayload rowChange4 = new RawTripTestPayload(recordStr4);
HoodieRecord record4 =
new HoodieRecord(new HoodieKey(rowChange4.getRowKey(), rowChange4.getPartitionPath()), rowChange4);
new HoodieAvroRecord(new HoodieKey(rowChange4.getRowKey(), rowChange4.getPartitionPath()), rowChange4);
List<HoodieRecord> records = asList(record1, record2, record3, record4);
// Also create the metadata and config
@@ -354,15 +355,15 @@ public class TestFlinkHoodieBloomIndex extends HoodieFlinkClientTestHarness {
+ "\"time\":\"2015-01-31T03:16:41.415Z\",\"number\":32}";
RawTripTestPayload rowChange1 = new RawTripTestPayload(recordStr1);
HoodieKey key1 = new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath());
HoodieRecord record1 = new HoodieRecord(key1, rowChange1);
HoodieRecord record1 = new HoodieAvroRecord(key1, rowChange1);
RawTripTestPayload rowChange2 = new RawTripTestPayload(recordStr2);
HoodieKey key2 = new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath());
HoodieRecord record2 = new HoodieRecord(key2, rowChange2);
HoodieRecord record2 = new HoodieAvroRecord(key2, rowChange2);
RawTripTestPayload rowChange3 = new RawTripTestPayload(recordStr3);
HoodieKey key3 = new HoodieKey(rowChange3.getRowKey(), rowChange3.getPartitionPath());
RawTripTestPayload rowChange4 = new RawTripTestPayload(recordStr4);
HoodieKey key4 = new HoodieKey(rowChange4.getRowKey(), rowChange4.getPartitionPath());
HoodieRecord record4 = new HoodieRecord(key4, rowChange4);
HoodieRecord record4 = new HoodieAvroRecord(key4, rowChange4);
List<HoodieKey> keys = asList(key1, key2, key3, key4);
// Also create the metadata and config
@@ -373,7 +374,7 @@ public class TestFlinkHoodieBloomIndex extends HoodieFlinkClientTestHarness {
// Let's tag
HoodieBloomIndex bloomIndex = new HoodieBloomIndex(config, ListBasedHoodieBloomIndexHelper.getInstance());
List<HoodieRecord> toTagRecords = new ArrayList<>();
toTagRecords.add(new HoodieRecord(record4.getKey(), null));
toTagRecords.add(new HoodieAvroRecord(record4.getKey(), null));
List<HoodieRecord> taggedRecords = tagLocation(bloomIndex, toTagRecords, hoodieTable);
Map<HoodieKey, Option<Pair<String, String>>> recordLocations = new HashMap<>();
for (HoodieRecord taggedRecord : taggedRecords) {
@@ -396,7 +397,7 @@ public class TestFlinkHoodieBloomIndex extends HoodieFlinkClientTestHarness {
hoodieTable = HoodieFlinkTable.create(config, context, metaClient);
List<HoodieRecord> toTagRecords1 = new ArrayList<>();
for (HoodieKey key : keys) {
taggedRecords.add(new HoodieRecord(key, null));
taggedRecords.add(new HoodieAvroRecord(key, null));
}
taggedRecords = tagLocation(bloomIndex, toTagRecords1, hoodieTable);
@@ -436,9 +437,9 @@ public class TestFlinkHoodieBloomIndex extends HoodieFlinkClientTestHarness {
// We write record1 to a base file, using a bloom filter having both records
RawTripTestPayload rowChange1 = new RawTripTestPayload(recordStr1);
HoodieRecord record1 = new HoodieRecord(new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath()), rowChange1);
HoodieRecord record1 = new HoodieAvroRecord(new HoodieKey(rowChange1.getRowKey(), rowChange1.getPartitionPath()), rowChange1);
RawTripTestPayload rowChange2 = new RawTripTestPayload(recordStr2);
HoodieRecord record2 = new HoodieRecord(new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath()), rowChange2);
HoodieRecord record2 = new HoodieAvroRecord(new HoodieKey(rowChange2.getRowKey(), rowChange2.getPartitionPath()), rowChange2);
BloomFilter filter = BloomFilterFactory.createBloomFilter(10000, 0.0000001, -1, BloomFilterTypeCode.SIMPLE.name());
filter.add(record2.getRecordKey());

3
hudi-client/hudi-flink-client/src/test/java/org/apache/hudi/testutils/HoodieFlinkWriteableTestTable.java
View File

@@ -27,6 +27,7 @@ import org.apache.hudi.common.bloom.BloomFilterTypeCode;
import org.apache.hudi.common.model.HoodieLogFile;
import org.apache.hudi.common.model.HoodieRecord;
import org.apache.hudi.common.model.HoodieRecordLocation;
import org.apache.hudi.common.model.HoodieRecordPayload;
import org.apache.hudi.common.table.HoodieTableMetaClient;
import org.apache.hudi.common.table.log.HoodieLogFormat;
import org.apache.hudi.common.table.log.block.HoodieAvroDataBlock;
@@ -132,7 +133,7 @@ public class HoodieFlinkWriteableTestTable extends HoodieWriteableTestTable {
header.put(HeaderMetadataType.SCHEMA, schema.toString());
logWriter.appendBlock(new HoodieAvroDataBlock(groupedRecords.stream().map(r -> {
try {
GenericRecord val = (GenericRecord) r.getData().getInsertValue(schema).get();
GenericRecord val = (GenericRecord) ((HoodieRecordPayload) r.getData()).getInsertValue(schema).get();
HoodieAvroUtils.addHoodieKeyToRecord(val, r.getRecordKey(), r.getPartitionPath(), "");
return (IndexedRecord) val;
} catch (IOException e) {