Search in sources :

Example 6 with AcidTable

use of org.apache.hive.hcatalog.streaming.mutate.client.AcidTable in project hive by apache.

the class TestMutations method testTransactionBatchCommitPartitioned.

@Test
public void testTransactionBatchCommitPartitioned() throws Exception {
    Table table = partitionedTableBuilder.addPartition(ASIA_INDIA).create(metaStoreClient);
    MutatorClient client = new MutatorClientBuilder().addSinkTable(table.getDbName(), table.getTableName(), true).metaStoreUri(metaStoreUri).build();
    client.connect();
    Transaction transaction = client.newTransaction();
    List<AcidTable> destinations = client.getTables();
    transaction.begin();
    MutatorFactory mutatorFactory = new ReflectiveMutatorFactory(conf, MutableRecord.class, RECORD_ID_COLUMN, BUCKET_COLUMN_INDEXES);
    MutatorCoordinator coordinator = new MutatorCoordinatorBuilder().metaStoreUri(metaStoreUri).table(destinations.get(0)).mutatorFactory(mutatorFactory).build();
    BucketIdResolver bucketIdAppender = mutatorFactory.newBucketIdResolver(destinations.get(0).getTotalBuckets());
    MutableRecord record = (MutableRecord) bucketIdAppender.attachBucketIdToRecord(new MutableRecord(1, "Hello streaming"));
    coordinator.insert(ASIA_INDIA, record);
    coordinator.close();
    transaction.commit();
    StreamingAssert streamingAssertions = assertionFactory.newStreamingAssert(table, ASIA_INDIA);
    streamingAssertions.assertMinWriteId(1L);
    streamingAssertions.assertMaxWriteId(1L);
    streamingAssertions.assertExpectedFileCount(1);
    List<Record> readRecords = streamingAssertions.readRecords();
    assertThat(readRecords.size(), is(1));
    assertThat(readRecords.get(0).getRow(), is("{1, Hello streaming}"));
    assertThat(readRecords.get(0).getRecordIdentifier(), is(new RecordIdentifier(1L, encodeBucket(0), 0L)));
    assertThat(transaction.getState(), is(COMMITTED));
    client.close();
}
Also used : AcidTable(org.apache.hive.hcatalog.streaming.mutate.client.AcidTable) Table(org.apache.hadoop.hive.metastore.api.Table) AcidTable(org.apache.hive.hcatalog.streaming.mutate.client.AcidTable) MutatorCoordinator(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorCoordinator) RecordIdentifier(org.apache.hadoop.hive.ql.io.RecordIdentifier) MutatorFactory(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorFactory) Transaction(org.apache.hive.hcatalog.streaming.mutate.client.Transaction) MutatorCoordinatorBuilder(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorCoordinatorBuilder) BucketIdResolver(org.apache.hive.hcatalog.streaming.mutate.worker.BucketIdResolver) Record(org.apache.hive.hcatalog.streaming.mutate.StreamingAssert.Record) MutatorClient(org.apache.hive.hcatalog.streaming.mutate.client.MutatorClient) MutatorClientBuilder(org.apache.hive.hcatalog.streaming.mutate.client.MutatorClientBuilder) Test(org.junit.Test)

Example 7 with AcidTable

use of org.apache.hive.hcatalog.streaming.mutate.client.AcidTable in project hive by apache.

the class TestMutations method testTransactionBatchCommitUnpartitioned.

@Test
public void testTransactionBatchCommitUnpartitioned() throws Exception {
    Table table = unpartitionedTableBuilder.create(metaStoreClient);
    MutatorClient client = new MutatorClientBuilder().addSinkTable(table.getDbName(), table.getTableName(), false).metaStoreUri(metaStoreUri).build();
    client.connect();
    Transaction transaction = client.newTransaction();
    List<AcidTable> destinations = client.getTables();
    transaction.begin();
    MutatorFactory mutatorFactory = new ReflectiveMutatorFactory(conf, MutableRecord.class, RECORD_ID_COLUMN, BUCKET_COLUMN_INDEXES);
    MutatorCoordinator coordinator = new MutatorCoordinatorBuilder().metaStoreUri(metaStoreUri).table(destinations.get(0)).mutatorFactory(mutatorFactory).build();
    BucketIdResolver bucketIdResolver = mutatorFactory.newBucketIdResolver(destinations.get(0).getTotalBuckets());
    MutableRecord record = (MutableRecord) bucketIdResolver.attachBucketIdToRecord(new MutableRecord(1, "Hello streaming"));
    coordinator.insert(Collections.<String>emptyList(), record);
    coordinator.close();
    transaction.commit();
    StreamingAssert streamingAssertions = assertionFactory.newStreamingAssert(table);
    streamingAssertions.assertMinWriteId(1L);
    streamingAssertions.assertMaxWriteId(1L);
    streamingAssertions.assertExpectedFileCount(1);
    List<Record> readRecords = streamingAssertions.readRecords();
    assertThat(readRecords.size(), is(1));
    assertThat(readRecords.get(0).getRow(), is("{1, Hello streaming}"));
    assertThat(readRecords.get(0).getRecordIdentifier(), is(new RecordIdentifier(1L, encodeBucket(0), 0L)));
    assertThat(transaction.getState(), is(COMMITTED));
    client.close();
}
Also used : AcidTable(org.apache.hive.hcatalog.streaming.mutate.client.AcidTable) Table(org.apache.hadoop.hive.metastore.api.Table) AcidTable(org.apache.hive.hcatalog.streaming.mutate.client.AcidTable) MutatorCoordinator(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorCoordinator) RecordIdentifier(org.apache.hadoop.hive.ql.io.RecordIdentifier) MutatorFactory(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorFactory) Transaction(org.apache.hive.hcatalog.streaming.mutate.client.Transaction) MutatorCoordinatorBuilder(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorCoordinatorBuilder) BucketIdResolver(org.apache.hive.hcatalog.streaming.mutate.worker.BucketIdResolver) Record(org.apache.hive.hcatalog.streaming.mutate.StreamingAssert.Record) MutatorClient(org.apache.hive.hcatalog.streaming.mutate.client.MutatorClient) MutatorClientBuilder(org.apache.hive.hcatalog.streaming.mutate.client.MutatorClientBuilder) Test(org.junit.Test)

Example 8 with AcidTable

use of org.apache.hive.hcatalog.streaming.mutate.client.AcidTable in project hive by apache.

the class TestMutations method testUpdatesAndDeletes.

@Test
public void testUpdatesAndDeletes() throws Exception {
    // Set up some base data then stream some inserts/updates/deletes to a number of partitions
    MutatorFactory mutatorFactory = new ReflectiveMutatorFactory(conf, MutableRecord.class, RECORD_ID_COLUMN, BUCKET_COLUMN_INDEXES);
    // INSERT DATA
    // 
    Table table = partitionedTableBuilder.addPartition(ASIA_INDIA).addPartition(EUROPE_FRANCE).create(metaStoreClient);
    MutatorClient client = new MutatorClientBuilder().addSinkTable(table.getDbName(), table.getTableName(), true).metaStoreUri(metaStoreUri).build();
    client.connect();
    Transaction insertTransaction = client.newTransaction();
    List<AcidTable> destinations = client.getTables();
    insertTransaction.begin();
    MutatorCoordinator insertCoordinator = new MutatorCoordinatorBuilder().metaStoreUri(metaStoreUri).table(destinations.get(0)).mutatorFactory(mutatorFactory).build();
    BucketIdResolver bucketIdResolver = mutatorFactory.newBucketIdResolver(destinations.get(0).getTotalBuckets());
    MutableRecord asiaIndiaRecord1 = (MutableRecord) bucketIdResolver.attachBucketIdToRecord(new MutableRecord(1, "Namaste streaming 1"));
    MutableRecord asiaIndiaRecord2 = (MutableRecord) bucketIdResolver.attachBucketIdToRecord(new MutableRecord(2, "Namaste streaming 2"));
    MutableRecord europeUkRecord1 = (MutableRecord) bucketIdResolver.attachBucketIdToRecord(new MutableRecord(3, "Hello streaming 1"));
    MutableRecord europeUkRecord2 = (MutableRecord) bucketIdResolver.attachBucketIdToRecord(new MutableRecord(4, "Hello streaming 2"));
    MutableRecord europeFranceRecord1 = (MutableRecord) bucketIdResolver.attachBucketIdToRecord(new MutableRecord(5, "Bonjour streaming 1"));
    MutableRecord europeFranceRecord2 = (MutableRecord) bucketIdResolver.attachBucketIdToRecord(new MutableRecord(6, "Bonjour streaming 2"));
    insertCoordinator.insert(ASIA_INDIA, asiaIndiaRecord1);
    insertCoordinator.insert(ASIA_INDIA, asiaIndiaRecord2);
    insertCoordinator.insert(EUROPE_UK, europeUkRecord1);
    insertCoordinator.insert(EUROPE_UK, europeUkRecord2);
    insertCoordinator.insert(EUROPE_FRANCE, europeFranceRecord1);
    insertCoordinator.insert(EUROPE_FRANCE, europeFranceRecord2);
    insertCoordinator.close();
    insertTransaction.commit();
    assertThat(insertTransaction.getState(), is(COMMITTED));
    client.close();
    // MUTATE DATA
    // 
    client = new MutatorClientBuilder().addSinkTable(table.getDbName(), table.getTableName(), true).metaStoreUri(metaStoreUri).build();
    client.connect();
    Transaction mutateTransaction = client.newTransaction();
    destinations = client.getTables();
    mutateTransaction.begin();
    MutatorCoordinator mutateCoordinator = new MutatorCoordinatorBuilder().metaStoreUri(metaStoreUri).table(destinations.get(0)).mutatorFactory(mutatorFactory).build();
    bucketIdResolver = mutatorFactory.newBucketIdResolver(destinations.get(0).getTotalBuckets());
    MutableRecord asiaIndiaRecord3 = (MutableRecord) bucketIdResolver.attachBucketIdToRecord(new MutableRecord(20, "Namaste streaming 3"));
    mutateCoordinator.update(ASIA_INDIA, new MutableRecord(2, "UPDATED: Namaste streaming 2", new RecordIdentifier(1L, encodeBucket(0), 1L)));
    mutateCoordinator.insert(ASIA_INDIA, asiaIndiaRecord3);
    mutateCoordinator.delete(EUROPE_UK, new MutableRecord(3, "Hello streaming 1", new RecordIdentifier(1L, encodeBucket(0), 0L)));
    mutateCoordinator.delete(EUROPE_FRANCE, new MutableRecord(5, "Bonjour streaming 1", new RecordIdentifier(1L, encodeBucket(0), 0L)));
    mutateCoordinator.update(EUROPE_FRANCE, new MutableRecord(6, "UPDATED: Bonjour streaming 2", new RecordIdentifier(1L, encodeBucket(0), 1L)));
    mutateCoordinator.close();
    mutateTransaction.commit();
    assertThat(mutateTransaction.getState(), is(COMMITTED));
    StreamingAssert indiaAssertions = assertionFactory.newStreamingAssert(table, ASIA_INDIA);
    indiaAssertions.assertMinWriteId(1L);
    indiaAssertions.assertMaxWriteId(2L);
    List<Record> indiaRecords = indiaAssertions.readRecords(2);
    assertThat(indiaRecords.size(), is(3));
    assertThat(indiaRecords.get(0).getRow(), is("{1, Namaste streaming 1}"));
    assertThat(indiaRecords.get(0).getRecordIdentifier(), is(new RecordIdentifier(1L, encodeBucket(0), 0L)));
    assertThat(indiaRecords.get(1).getRow(), is("{2, UPDATED: Namaste streaming 2}"));
    assertThat(indiaRecords.get(1).getRecordIdentifier(), is(new RecordIdentifier(2L, encodeBucket(0), // with split update, new version of the row is a new insert
    0L)));
    assertThat(indiaRecords.get(2).getRow(), is("{20, Namaste streaming 3}"));
    assertThat(indiaRecords.get(2).getRecordIdentifier(), is(new RecordIdentifier(2L, encodeBucket(0), 1L)));
    StreamingAssert ukAssertions = assertionFactory.newStreamingAssert(table, EUROPE_UK);
    ukAssertions.assertMinWriteId(1L);
    ukAssertions.assertMaxWriteId(2L);
    // 1 split since mutateTransaction txn just does deletes
    List<Record> ukRecords = ukAssertions.readRecords(1);
    assertThat(ukRecords.size(), is(1));
    assertThat(ukRecords.get(0).getRow(), is("{4, Hello streaming 2}"));
    assertThat(ukRecords.get(0).getRecordIdentifier(), is(new RecordIdentifier(1L, encodeBucket(0), 1L)));
    StreamingAssert franceAssertions = assertionFactory.newStreamingAssert(table, EUROPE_FRANCE);
    franceAssertions.assertMinWriteId(1L);
    franceAssertions.assertMaxWriteId(2L);
    List<Record> franceRecords = franceAssertions.readRecords(2);
    assertThat(franceRecords.size(), is(1));
    assertThat(franceRecords.get(0).getRow(), is("{6, UPDATED: Bonjour streaming 2}"));
    assertThat(franceRecords.get(0).getRecordIdentifier(), is(new RecordIdentifier(2L, encodeBucket(0), // with split update, new version of the row is a new insert
    0L)));
    client.close();
}
Also used : AcidTable(org.apache.hive.hcatalog.streaming.mutate.client.AcidTable) Table(org.apache.hadoop.hive.metastore.api.Table) AcidTable(org.apache.hive.hcatalog.streaming.mutate.client.AcidTable) MutatorCoordinator(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorCoordinator) RecordIdentifier(org.apache.hadoop.hive.ql.io.RecordIdentifier) MutatorFactory(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorFactory) Transaction(org.apache.hive.hcatalog.streaming.mutate.client.Transaction) MutatorCoordinatorBuilder(org.apache.hive.hcatalog.streaming.mutate.worker.MutatorCoordinatorBuilder) BucketIdResolver(org.apache.hive.hcatalog.streaming.mutate.worker.BucketIdResolver) Record(org.apache.hive.hcatalog.streaming.mutate.StreamingAssert.Record) MutatorClient(org.apache.hive.hcatalog.streaming.mutate.client.MutatorClient) MutatorClientBuilder(org.apache.hive.hcatalog.streaming.mutate.client.MutatorClientBuilder) Test(org.junit.Test)

Aggregations

AcidTable (org.apache.hive.hcatalog.streaming.mutate.client.AcidTable)8 MutatorClient (org.apache.hive.hcatalog.streaming.mutate.client.MutatorClient)8 MutatorClientBuilder (org.apache.hive.hcatalog.streaming.mutate.client.MutatorClientBuilder)8 Transaction (org.apache.hive.hcatalog.streaming.mutate.client.Transaction)8 MutatorCoordinator (org.apache.hive.hcatalog.streaming.mutate.worker.MutatorCoordinator)8 MutatorCoordinatorBuilder (org.apache.hive.hcatalog.streaming.mutate.worker.MutatorCoordinatorBuilder)8 Table (org.apache.hadoop.hive.metastore.api.Table)7 MutatorFactory (org.apache.hive.hcatalog.streaming.mutate.worker.MutatorFactory)7 Test (org.junit.Test)7 BucketIdResolver (org.apache.hive.hcatalog.streaming.mutate.worker.BucketIdResolver)6 RecordIdentifier (org.apache.hadoop.hive.ql.io.RecordIdentifier)4 Record (org.apache.hive.hcatalog.streaming.mutate.StreamingAssert.Record)4