use of org.apache.drill.metastore.metadata.TableInfo in project drill by apache.
the class TestTableMetadataUnitConversion method testSegmentMetadata.
@Test
public void testSegmentMetadata() {
TableInfo tableInfo = data.basicTableInfo;
MetadataInfo metadataInfo = MetadataInfo.builder().type(MetadataType.SEGMENT).key("part_int=3").identifier("part_int=3/part_varchar=g").build();
Path path = new Path("/tmp/nation");
String unitPath = path.toUri().getPath();
Set<Path> locations = new HashSet<>();
locations.add(new Path("part_int=3/part_varchar=g/0_0_0.parquet"));
locations.add(new Path("part_int=3/part_varchar=g/0_0_1.parquet"));
List<String> unitLocations = locations.stream().map(location -> location.toUri().getPath()).collect(Collectors.toList());
// check required fields
SegmentMetadata requiredFieldsMetadata = SegmentMetadata.builder().tableInfo(tableInfo).metadataInfo(metadataInfo).columnsStatistics(data.columnsStatistics).metadataStatistics(data.metadataStatistics).path(path).locations(locations).build();
TableMetadataUnit requiredFieldsExpectedUnit = TableMetadataUnit.builder().storagePlugin(tableInfo.storagePlugin()).workspace(tableInfo.workspace()).tableName(tableInfo.name()).metadataType(metadataInfo.type().name()).metadataKey(metadataInfo.key()).metadataIdentifier(metadataInfo.identifier()).columnsStatistics(data.unitColumnsStatistics).metadataStatistics(data.unitMetadataStatistics).lastModifiedTime(BaseMetadata.UNDEFINED_TIME).path(path.toUri().getPath()).location(unitPath).locations(unitLocations).build();
TableMetadataUnit requiredFieldsUnit = requiredFieldsMetadata.toMetadataUnit();
assertEquals(requiredFieldsExpectedUnit, requiredFieldsUnit);
assertNotNull(SegmentMetadata.builder().metadataUnit(requiredFieldsUnit).build());
SchemaPath column = SchemaPath.getSimplePath("dir1");
List<String> partitionValues = Collections.singletonList("part_varchar=g");
SegmentMetadata allFieldsMetadata = SegmentMetadata.builder().tableInfo(tableInfo).metadataInfo(metadataInfo).schema(data.schema).columnsStatistics(data.columnsStatistics).metadataStatistics(data.metadataStatistics).lastModifiedTime(data.lastModifiedTime).path(path).locations(locations).column(column).partitionValues(partitionValues).build();
TableMetadataUnit allFieldsExpectedUnit = TableMetadataUnit.builder().storagePlugin(tableInfo.storagePlugin()).workspace(tableInfo.workspace()).tableName(tableInfo.name()).metadataType(metadataInfo.type().name()).metadataKey(metadataInfo.key()).metadataIdentifier(metadataInfo.identifier()).schema(data.unitSchema).columnsStatistics(data.unitColumnsStatistics).metadataStatistics(data.unitMetadataStatistics).lastModifiedTime(data.lastModifiedTime).path(path.toUri().getPath()).location(unitPath).locations(unitLocations).column(column.toString()).partitionValues(partitionValues).build();
TableMetadataUnit allFieldsUnit = allFieldsMetadata.toMetadataUnit();
assertEquals(allFieldsExpectedUnit, allFieldsUnit);
assertNotNull(SegmentMetadata.builder().metadataUnit(allFieldsUnit).build());
}
use of org.apache.drill.metastore.metadata.TableInfo in project drill by apache.
the class TestMetastoreWithEasyFormatPlugin method testNonEmptyTableWithEmptyFile.
@Test
public void testNonEmptyTableWithEmptyFile() throws Exception {
String tableName = "csv_with_empty_file";
dirTestWatcher.copyResourceToTestTmp(Paths.get("store", "text", "directoryWithEmptyCSV", "empty.csv"), Paths.get(tableName, "empty.csv"));
File table = dirTestWatcher.copyResourceToTestTmp(Paths.get("store", "text", "data", "nations.csv"), Paths.get(tableName, "nations.csv")).getParentFile();
TableInfo tableInfo = getTableInfo(tableName, "tmp", "csv");
TupleMetadata schema = new SchemaBuilder().add("n_nationkey", TypeProtos.MinorType.INT).add("n_name", TypeProtos.MinorType.VARCHAR).add("n_regionkey", TypeProtos.MinorType.INT).add("n_comment", TypeProtos.MinorType.VARCHAR).build();
Map<SchemaPath, ColumnStatistics<?>> columnStatistics = ImmutableMap.<SchemaPath, ColumnStatistics<?>>builder().put(SchemaPath.getSimplePath("n_nationkey"), getColumnStatistics(0, 24, 25L, TypeProtos.MinorType.INT)).put(SchemaPath.getSimplePath("n_name"), getColumnStatistics("ALGERIA", "VIETNAM", 25L, TypeProtos.MinorType.VARCHAR)).put(SchemaPath.getSimplePath("n_regionkey"), getColumnStatistics(0, 4, 25L, TypeProtos.MinorType.INT)).put(SchemaPath.getSimplePath("n_comment"), getColumnStatistics("alfoxespromiseslylyaccordingtotheregularaccounts.boldrequestsalon", "yfinalpackages.slowfoxescajolequickly.quicklysilentplateletsbreachironicaccounts.unusualpintobe", 25L, TypeProtos.MinorType.VARCHAR)).build();
BaseTableMetadata expectedTableMetadata = BaseTableMetadata.builder().tableInfo(tableInfo).metadataInfo(TABLE_META_INFO).schema(schema).location(new Path(table.toURI().getPath())).columnsStatistics(columnStatistics).metadataStatistics(Arrays.asList(new StatisticsHolder<>(25L, TableStatisticsKind.ROW_COUNT), new StatisticsHolder<>(MetadataType.ALL, TableStatisticsKind.ANALYZE_METADATA_LEVEL))).partitionKeys(Collections.emptyMap()).lastModifiedTime(getMaxLastModified(table)).build();
try {
testBuilder().sqlQuery("ANALYZE TABLE table(dfs.tmp.`%s` " + "(schema=>'inline=(" + "`n_nationkey` INT not null," + "`n_name` VARCHAR not null," + "`n_regionkey` INT not null," + "`n_comment` VARCHAR not null)')) REFRESH METADATA", tableName).unOrdered().baselineColumns("ok", "summary").baselineValues(true, String.format("Collected / refreshed metadata for table [dfs.tmp.%s]", tableName)).go();
MetastoreTableInfo metastoreTableInfo = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().metastoreTableInfo(tableInfo);
assertTrue("table metadata wasn't found", metastoreTableInfo.isExists());
BaseTableMetadata tableMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().tableMetadata(tableInfo);
assertEquals(expectedTableMetadata, tableMetadata);
List<FileMetadata> filesMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().filesMetadata(tableInfo, null, null);
assertEquals(2, filesMetadata.size());
String query = "select * from dfs.tmp.`%s`";
queryBuilder().sql(query, tableName).planMatcher().include("usedMetastore=true").match();
long rowCount = queryBuilder().sql(query, tableName).run().recordCount();
assertEquals(25, rowCount);
} finally {
run("analyze table dfs.tmp.`%s` drop metadata if exists", tableName);
}
}
use of org.apache.drill.metastore.metadata.TableInfo in project drill by apache.
the class TestMetastoreWithEasyFormatPlugin method testAnalyzeOnTextTableWithHeader.
@Test
public void testAnalyzeOnTextTableWithHeader() throws Exception {
String tableName = "store/text/data/cars.csvh";
File tablePath = dirTestWatcher.copyResourceToRoot(Paths.get(tableName));
TableInfo tableInfo = getTableInfo(tableName, "default", "csvh");
TupleMetadata schema = new SchemaBuilder().add("Year", TypeProtos.MinorType.VARCHAR).add("Make", TypeProtos.MinorType.VARCHAR).add("Model", TypeProtos.MinorType.VARCHAR).add("Description", TypeProtos.MinorType.VARCHAR).add("Price", TypeProtos.MinorType.VARCHAR).build();
ImmutableMap<SchemaPath, ColumnStatistics<?>> tableColumnStatistics = ImmutableMap.<SchemaPath, ColumnStatistics<?>>builder().put(SchemaPath.getSimplePath("Description"), getColumnStatistics("", "ac, abs, moon", 4L, TypeProtos.MinorType.VARCHAR)).put(SchemaPath.getSimplePath("Make"), getColumnStatistics("Chevy", "Jeep", 4L, TypeProtos.MinorType.VARCHAR)).put(SchemaPath.getSimplePath("Model"), getColumnStatistics("E350", "Venture \"Extended Edition, Very Large\"", 4L, TypeProtos.MinorType.VARCHAR)).put(SchemaPath.getSimplePath("Price"), getColumnStatistics("3000.00", "5000.00", 4L, TypeProtos.MinorType.VARCHAR)).put(SchemaPath.getSimplePath("Year"), getColumnStatistics("1996", "1999", 4L, TypeProtos.MinorType.VARCHAR)).build();
BaseTableMetadata expectedTableMetadata = BaseTableMetadata.builder().tableInfo(tableInfo).metadataInfo(TABLE_META_INFO).schema(schema).location(new Path(tablePath.toURI().getPath())).columnsStatistics(tableColumnStatistics).metadataStatistics(Arrays.asList(new StatisticsHolder<>(4L, TableStatisticsKind.ROW_COUNT), new StatisticsHolder<>(MetadataType.ALL, TableStatisticsKind.ANALYZE_METADATA_LEVEL))).partitionKeys(Collections.emptyMap()).lastModifiedTime(getMaxLastModified(tablePath)).build();
try {
testBuilder().sqlQuery("analyze table dfs.`%s` refresh metadata", tableName).unOrdered().baselineColumns("ok", "summary").baselineValues(true, String.format("Collected / refreshed metadata for table [dfs.default.%s]", tableName)).go();
BaseTableMetadata actualTableMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().tableMetadata(tableInfo);
assertEquals(expectedTableMetadata, actualTableMetadata);
} finally {
run("analyze table dfs.`%s` drop metadata if exists", tableName);
}
}
use of org.apache.drill.metastore.metadata.TableInfo in project drill by apache.
the class TestMetastoreWithEasyFormatPlugin method testAnalyzeOnJsonTable.
@Test
public void testAnalyzeOnJsonTable() throws Exception {
String tableName = "multilevel/json";
TableInfo tableInfo = getTableInfo(tableName, "default", "json");
File table = dirTestWatcher.copyResourceToRoot(Paths.get(tableName));
Path tablePath = new Path(table.toURI().getPath());
TupleMetadata schema = new SchemaBuilder().addNullable("dir0", TypeProtos.MinorType.VARCHAR).addNullable("dir1", TypeProtos.MinorType.VARCHAR).addNullable("o_orderkey", TypeProtos.MinorType.BIGINT).addNullable("o_custkey", TypeProtos.MinorType.BIGINT).addNullable("o_orderstatus", TypeProtos.MinorType.VARCHAR).addNullable("o_totalprice", TypeProtos.MinorType.FLOAT8).addNullable("o_orderdate", TypeProtos.MinorType.VARCHAR).addNullable("o_orderpriority", TypeProtos.MinorType.VARCHAR).addNullable("o_clerk", TypeProtos.MinorType.VARCHAR).addNullable("o_shippriority", TypeProtos.MinorType.BIGINT).addNullable("o_comment", TypeProtos.MinorType.VARCHAR).build();
Map<SchemaPath, ColumnStatistics<?>> tableColumnStatistics = new HashMap<>(TABLE_COLUMN_STATISTICS);
tableColumnStatistics.put(SchemaPath.getSimplePath("o_custkey"), getColumnStatistics(25L, 1498L, 120L, TypeProtos.MinorType.BIGINT));
tableColumnStatistics.put(SchemaPath.getSimplePath("o_orderdate"), getColumnStatistics("1994-01-01T00:00:00.000-08:00", "1996-12-19T00:00:00.000-08:00", 120L, TypeProtos.MinorType.VARCHAR));
tableColumnStatistics.put(SchemaPath.getSimplePath("o_orderkey"), getColumnStatistics(1L, 1319L, 120L, TypeProtos.MinorType.BIGINT));
tableColumnStatistics.put(SchemaPath.getSimplePath("o_shippriority"), getColumnStatistics(0L, 0L, 120L, TypeProtos.MinorType.BIGINT));
BaseTableMetadata expectedTableMetadata = BaseTableMetadata.builder().tableInfo(tableInfo).metadataInfo(TABLE_META_INFO).schema(schema).location(new Path(table.toURI().getPath())).columnsStatistics(tableColumnStatistics).metadataStatistics(Arrays.asList(new StatisticsHolder<>(120L, TableStatisticsKind.ROW_COUNT), new StatisticsHolder<>(MetadataType.ALL, TableStatisticsKind.ANALYZE_METADATA_LEVEL))).partitionKeys(Collections.emptyMap()).lastModifiedTime(getMaxLastModified(table)).build();
TableInfo baseTableInfo = TableInfo.builder().name(tableName).storagePlugin("dfs").workspace("default").build();
Map<SchemaPath, ColumnStatistics<?>> dir0CSVStats = new HashMap<>(DIR0_1994_SEGMENT_COLUMN_STATISTICS);
dir0CSVStats.put(SchemaPath.getSimplePath("o_custkey"), getColumnStatistics(25L, 1469L, 40L, TypeProtos.MinorType.BIGINT));
dir0CSVStats.put(SchemaPath.getSimplePath("o_orderdate"), getColumnStatistics("1994-01-01T00:00:00.000-08:00", "1994-12-23T00:00:00.000-08:00", 40L, TypeProtos.MinorType.VARCHAR));
dir0CSVStats.put(SchemaPath.getSimplePath("o_orderkey"), getColumnStatistics(5L, 1031L, 40L, TypeProtos.MinorType.BIGINT));
dir0CSVStats.put(SchemaPath.getSimplePath("o_shippriority"), getColumnStatistics(0L, 0L, 40L, TypeProtos.MinorType.BIGINT));
SegmentMetadata dir0 = SegmentMetadata.builder().tableInfo(baseTableInfo).metadataInfo(MetadataInfo.builder().type(MetadataType.SEGMENT).identifier("1994").key("1994").build()).path(new Path(tablePath, "1994")).schema(schema).lastModifiedTime(getMaxLastModified(new File(table, "1994"))).column(SchemaPath.getSimplePath("dir0")).columnsStatistics(dir0CSVStats).metadataStatistics(Collections.singletonList(new StatisticsHolder<>(40L, TableStatisticsKind.ROW_COUNT))).locations(ImmutableSet.of(new Path(tablePath, "1994/Q1/orders_94_q1.json"), new Path(tablePath, "1994/Q2/orders_94_q2.json"), new Path(tablePath, "1994/Q3/orders_94_q3.json"), new Path(tablePath, "1994/Q4/orders_94_q4.json"))).partitionValues(Collections.singletonList("1994")).build();
Set<Path> expectedTopLevelSegmentLocations = ImmutableSet.of(new Path(tablePath, "1994"), new Path(tablePath, "1995"), new Path(tablePath, "1996"));
Set<Set<Path>> expectedSegmentFilesLocations = new HashSet<>();
Set<Path> segmentFiles = ImmutableSet.of(new Path(tablePath, "1994/Q2/orders_94_q2.json"), new Path(tablePath, "1994/Q4/orders_94_q4.json"), new Path(tablePath, "1994/Q1/orders_94_q1.json"), new Path(tablePath, "1994/Q3/orders_94_q3.json"));
expectedSegmentFilesLocations.add(segmentFiles);
segmentFiles = ImmutableSet.of(new Path(tablePath, "1995/Q2/orders_95_q2.json"), new Path(tablePath, "1995/Q4/orders_95_q4.json"), new Path(tablePath, "1995/Q1/orders_95_q1.json"), new Path(tablePath, "1995/Q3/orders_95_q3.json"));
expectedSegmentFilesLocations.add(segmentFiles);
segmentFiles = ImmutableSet.of(new Path(tablePath, "1996/Q3/orders_96_q3.json"), new Path(tablePath, "1996/Q2/orders_96_q2.json"), new Path(tablePath, "1996/Q4/orders_96_q4.json"), new Path(tablePath, "1996/Q1/orders_96_q1.json"));
expectedSegmentFilesLocations.add(segmentFiles);
Map<SchemaPath, ColumnStatistics<?>> dir0q1Stats = new HashMap<>(DIR0_1994_Q1_SEGMENT_COLUMN_STATISTICS);
dir0q1Stats.put(SchemaPath.getSimplePath("o_custkey"), getColumnStatistics(392L, 1411L, 10L, TypeProtos.MinorType.BIGINT));
dir0q1Stats.put(SchemaPath.getSimplePath("o_orderdate"), getColumnStatistics("1994-01-01T00:00:00.000-08:00", "1994-03-26T00:00:00.000-08:00", 10L, TypeProtos.MinorType.VARCHAR));
dir0q1Stats.put(SchemaPath.getSimplePath("o_orderkey"), getColumnStatistics(66L, 833L, 10L, TypeProtos.MinorType.BIGINT));
dir0q1Stats.put(SchemaPath.getSimplePath("o_shippriority"), getColumnStatistics(0L, 0L, 10L, TypeProtos.MinorType.BIGINT));
long dir0q1lastModified = new File(new File(new File(table, "1994"), "Q1"), "orders_94_q1.json").lastModified();
FileMetadata dir01994q1File = FileMetadata.builder().tableInfo(baseTableInfo).metadataInfo(MetadataInfo.builder().type(MetadataType.FILE).identifier("1994/Q1/orders_94_q1.json").key("1994").build()).schema(schema).lastModifiedTime(dir0q1lastModified).columnsStatistics(dir0q1Stats).metadataStatistics(Collections.singletonList(new StatisticsHolder<>(10L, TableStatisticsKind.ROW_COUNT))).path(new Path(tablePath, "1994/Q1/orders_94_q1.json")).build();
try {
testBuilder().sqlQuery("analyze table table(dfs.`%s`(schema=>%s)) refresh metadata", tableName, SCHEMA_STRING).unOrdered().baselineColumns("ok", "summary").baselineValues(true, String.format("Collected / refreshed metadata for table [dfs.default.%s]", tableName)).go();
BaseTableMetadata actualTableMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().tableMetadata(tableInfo);
assertEquals(expectedTableMetadata, actualTableMetadata);
List<SegmentMetadata> topSegmentMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().segmentsMetadataByColumn(tableInfo, null, "`dir0`");
SegmentMetadata actualDir0Metadata = topSegmentMetadata.stream().filter(unit -> unit.getMetadataInfo().identifier().equals("1994")).findAny().orElseThrow(() -> new AssertionError("Segment is absent"));
Set<Path> locations = actualDir0Metadata.getLocations();
actualDir0Metadata.toBuilder().locations(locations);
assertEquals(dir0, actualDir0Metadata);
Set<Path> topLevelSegmentLocations = topSegmentMetadata.stream().map(SegmentMetadata::getLocation).collect(Collectors.toSet());
// verify top segments locations
assertEquals(expectedTopLevelSegmentLocations, topLevelSegmentLocations);
Set<Set<Path>> segmentFilesLocations = topSegmentMetadata.stream().map(SegmentMetadata::getLocations).collect(Collectors.toSet());
assertEquals(expectedSegmentFilesLocations, segmentFilesLocations);
// verify nested segments
List<SegmentMetadata> nestedSegmentMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().segmentsMetadataByColumn(tableInfo, null, "`dir1`");
assertEquals(12, nestedSegmentMetadata.size());
SegmentMetadata dir01994q1Segment = SegmentMetadata.builder().tableInfo(baseTableInfo).metadataInfo(MetadataInfo.builder().type(MetadataType.SEGMENT).identifier("1994/Q1").key("1994").build()).path(new Path(new Path(tablePath, "1994"), "Q1")).schema(schema).lastModifiedTime(getMaxLastModified(new File(new File(table, "1994"), "Q1"))).column(SchemaPath.getSimplePath("dir1")).columnsStatistics(dir0q1Stats).metadataStatistics(Collections.singletonList(new StatisticsHolder<>(10L, TableStatisticsKind.ROW_COUNT))).locations(ImmutableSet.of(new Path(tablePath, "1994/Q1/orders_94_q1.json"))).partitionValues(Collections.singletonList("Q1")).build();
// verify segment for 1994
assertEquals(dir01994q1Segment, nestedSegmentMetadata.stream().filter(unit -> unit.getMetadataInfo().identifier().equals("1994/Q1")).findAny().orElse(null));
// verify files metadata
List<FileMetadata> filesMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().filesMetadata(tableInfo, null, null);
assertEquals(12, filesMetadata.size());
// verify first file metadata
assertEquals(dir01994q1File, filesMetadata.stream().filter(unit -> unit.getMetadataInfo().identifier().equals("1994/Q1/orders_94_q1.json")).findAny().orElse(null));
} finally {
run("analyze table dfs.`%s` drop metadata if exists", tableName);
}
}
use of org.apache.drill.metastore.metadata.TableInfo in project drill by apache.
the class TestMetastoreWithEasyFormatPlugin method testEmptyCSV.
@Test
public void testEmptyCSV() throws Exception {
String tableName = "store/text/directoryWithEmptyCSV/empty.csv";
File tablePath = dirTestWatcher.copyResourceToRoot(Paths.get(tableName));
TableInfo tableInfo = getTableInfo(tableName, "default", "csv");
TupleMetadata schema = new SchemaBuilder().add("Description", TypeProtos.MinorType.VARCHAR).build();
ImmutableMap<SchemaPath, ColumnStatistics<?>> tableColumnStatistics = ImmutableMap.<SchemaPath, ColumnStatistics<?>>builder().put(SchemaPath.getSimplePath("Description"), getColumnStatistics(null, null, 0L, TypeProtos.MinorType.VARCHAR)).build();
BaseTableMetadata expectedTableMetadata = BaseTableMetadata.builder().tableInfo(tableInfo).metadataInfo(TABLE_META_INFO).schema(schema).location(new Path(tablePath.toURI().getPath())).columnsStatistics(tableColumnStatistics).metadataStatistics(Arrays.asList(new StatisticsHolder<>(0L, TableStatisticsKind.ROW_COUNT), new StatisticsHolder<>(MetadataType.ALL, TableStatisticsKind.ANALYZE_METADATA_LEVEL))).partitionKeys(Collections.emptyMap()).lastModifiedTime(getMaxLastModified(tablePath)).build();
try {
testBuilder().sqlQuery("analyze table table(dfs.`%s` (schema=>'inline=(`Description` VARCHAR not null)')) refresh metadata", tableName).unOrdered().baselineColumns("ok", "summary").baselineValues(true, String.format("Collected / refreshed metadata for table [dfs.default.%s]", tableName)).go();
MetastoreTableInfo metastoreTableInfo = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().metastoreTableInfo(tableInfo);
assertTrue("table metadata wasn't found", metastoreTableInfo.isExists());
BaseTableMetadata tableMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().tableMetadata(tableInfo);
assertEquals(expectedTableMetadata, tableMetadata);
List<FileMetadata> filesMetadata = cluster.drillbit().getContext().getMetastoreRegistry().get().tables().basicRequests().filesMetadata(tableInfo, null, null);
assertEquals(1, filesMetadata.size());
String query = "select * from dfs.`%s`";
queryBuilder().sql(query, tableName).planMatcher().include("usedMetastore=true").match();
testBuilder().sqlQuery(query, tableName).unOrdered().baselineColumns("Description").expectsEmptyResultSet().go();
} finally {
run("analyze table dfs.`%s` drop metadata if exists", tableName);
}
}
Aggregations