Examples with CsvInputFormat org.apache.druid.data.input.impl.CsvInputFormat used on opensource projects

Example 16 with CsvInputFormat

use of org.apache.druid.data.input.impl.CsvInputFormat in project druid by druid-io.

the class IndexTaskTest method testMultipleParseExceptionsFailure.

@Test
public void testMultipleParseExceptionsFailure() throws Exception {
    final File tmpDir = temporaryFolder.newFolder();
    final File tmpFile = File.createTempFile("druid", "index", tmpDir);
    try (BufferedWriter writer = Files.newWriter(tmpFile, StandardCharsets.UTF_8)) {
        writer.write("time,dim,dimLong,dimFloat,val\n");
        // unparseable
        writer.write("unparseable,a,2,3.0,1\n");
        // valid row
        writer.write("2014-01-01T00:00:10Z,a,2,3.0,1\n");
        // unparseable
        writer.write("9.0,a,2,3.0,1\n");
        // thrown away
        writer.write("3014-03-01T00:00:10Z,outsideofinterval,2,3.0,1\n");
        // unparseable
        writer.write("99999999999-01-01T00:00:10Z,b,2,3.0,1\n");
    }
    // Allow up to 3 parse exceptions, and save up to 2 parse exceptions
    final IndexTuningConfig tuningConfig = new IndexTuningConfig(null, null, null, null, null, null, null, null, null, null, new DynamicPartitionsSpec(2, null), INDEX_SPEC, null, null, false, false, null, null, null, true, 2, 5, null, null);
    final TimestampSpec timestampSpec = new TimestampSpec("time", "auto", null);
    final DimensionsSpec dimensionsSpec = new DimensionsSpec(Arrays.asList(new StringDimensionSchema("dim"), new LongDimensionSchema("dimLong"), new FloatDimensionSchema("dimFloat")));
    final List<String> columns = Arrays.asList("time", "dim", "dimLong", "dimFloat", "val");
    final IndexIngestionSpec ingestionSpec;
    List<String> expectedMessages;
    if (useInputFormatApi) {
        ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, timestampSpec, dimensionsSpec, new CsvInputFormat(columns, null, null, true, 0), null, null, tuningConfig, false, false);
        expectedMessages = Arrays.asList(StringUtils.format("Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 3, Line: 6)", tmpFile.toURI()), StringUtils.format("Timestamp[9.0] is unparseable! Event: {time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 2, Line: 4)", tmpFile.toURI()), StringUtils.format("Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 1, Line: 2)", tmpFile.toURI()));
    } else {
        ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, new CSVParseSpec(timestampSpec, dimensionsSpec, null, columns, true, 0), null, null, tuningConfig, false, false);
        expectedMessages = Arrays.asList("Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "Timestamp[9.0] is unparseable! Event: {time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1}", "Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
    }
    IndexTask indexTask = new IndexTask(null, null, ingestionSpec, null);
    TaskStatus status = runTask(indexTask).lhs;
    Assert.assertEquals(TaskState.FAILED, status.getStatusCode());
    checkTaskStatusErrorMsgForParseExceptionsExceeded(status);
    IngestionStatsAndErrorsTaskReportData reportData = getTaskReportData();
    Map<String, Object> expectedMetrics = ImmutableMap.of(RowIngestionMeters.DETERMINE_PARTITIONS, ImmutableMap.of(RowIngestionMeters.PROCESSED_WITH_ERROR, 0, RowIngestionMeters.PROCESSED, 0, RowIngestionMeters.UNPARSEABLE, 0, RowIngestionMeters.THROWN_AWAY, 0), RowIngestionMeters.BUILD_SEGMENTS, ImmutableMap.of(RowIngestionMeters.PROCESSED_WITH_ERROR, 0, RowIngestionMeters.PROCESSED, 1, RowIngestionMeters.UNPARSEABLE, 3, RowIngestionMeters.THROWN_AWAY, useInputFormatApi ? 1 : 2));
    Assert.assertEquals(expectedMetrics, reportData.getRowStats());
    List<LinkedHashMap> parseExceptionReports = (List<LinkedHashMap>) reportData.getUnparseableEvents().get(RowIngestionMeters.BUILD_SEGMENTS);
    List<String> actualMessages = parseExceptionReports.stream().map((r) -> {
        return ((List<String>) r.get("details")).get(0);
    }).collect(Collectors.toList());
    Assert.assertEquals(expectedMessages, actualMessages);
    List<String> expectedInputs = Arrays.asList("{time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "{time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1}", "{time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
    List<String> actualInputs = parseExceptionReports.stream().map((r) -> {
        return (String) r.get("input");
    }).collect(Collectors.toList());
    Assert.assertEquals(expectedInputs, actualInputs);
}

Example 17 with CsvInputFormat

use of org.apache.druid.data.input.impl.CsvInputFormat in project druid by druid-io.

the class IndexTaskTest method testReportParseException.

@Test
public void testReportParseException() throws Exception {
    final File tmpDir = temporaryFolder.newFolder();
    final File tmpFile = File.createTempFile("druid", "index", tmpDir);
    try (BufferedWriter writer = Files.newWriter(tmpFile, StandardCharsets.UTF_8)) {
        writer.write("time,d,val\n");
        writer.write("unparseable,a,1\n");
        writer.write("2014-01-01T00:00:10Z,a,1\n");
    }
    final TimestampSpec timestampSpec = new TimestampSpec("time", "auto", null);
    final List<String> columns = Arrays.asList("time", "dim", "val");
    // report parse exception
    final IndexTuningConfig tuningConfig = createTuningConfig(2, null, null, null, null, false, true);
    final IndexIngestionSpec indexIngestionSpec;
    List<String> expectedMessages;
    if (useInputFormatApi) {
        indexIngestionSpec = createIngestionSpec(jsonMapper, tmpDir, timestampSpec, DimensionsSpec.EMPTY, new CsvInputFormat(columns, null, null, true, 0), null, null, tuningConfig, false, false);
        expectedMessages = ImmutableList.of(StringUtils.format("Timestamp[unparseable] is unparseable! Event: {time=unparseable, d=a, val=1} (Path: %s, Record: 1, Line: 2)", tmpFile.toURI()));
    } else {
        indexIngestionSpec = createIngestionSpec(jsonMapper, tmpDir, new CSVParseSpec(timestampSpec, DimensionsSpec.EMPTY, null, columns, true, 0), null, null, tuningConfig, false, false);
        expectedMessages = ImmutableList.of("Timestamp[unparseable] is unparseable! Event: {time=unparseable, d=a, val=1}");
    }
    IndexTask indexTask = new IndexTask(null, null, indexIngestionSpec, null);
    TaskStatus status = runTask(indexTask).lhs;
    Assert.assertEquals(TaskState.FAILED, status.getStatusCode());
    checkTaskStatusErrorMsgForParseExceptionsExceeded(status);
    IngestionStatsAndErrorsTaskReportData reportData = getTaskReportData();
    List<LinkedHashMap> parseExceptionReports = (List<LinkedHashMap>) reportData.getUnparseableEvents().get(RowIngestionMeters.BUILD_SEGMENTS);
    List<String> actualMessages = parseExceptionReports.stream().map((r) -> {
        return ((List<String>) r.get("details")).get(0);
    }).collect(Collectors.toList());
    Assert.assertEquals(expectedMessages, actualMessages);
    List<String> expectedInputs = ImmutableList.of("{time=unparseable, d=a, val=1}");
    List<String> actualInputs = parseExceptionReports.stream().map((r) -> {
        return (String) r.get("input");
    }).collect(Collectors.toList());
    Assert.assertEquals(expectedInputs, actualInputs);
}

Example 18 with CsvInputFormat

use of org.apache.druid.data.input.impl.CsvInputFormat in project druid by druid-io.

the class IndexTaskTest method testMultipleParseExceptionsFailureAtDeterminePartitions.

@Test
public void testMultipleParseExceptionsFailureAtDeterminePartitions() throws Exception {
    final File tmpDir = temporaryFolder.newFolder();
    final File tmpFile = File.createTempFile("druid", "index", tmpDir);
    try (BufferedWriter writer = Files.newWriter(tmpFile, StandardCharsets.UTF_8)) {
        writer.write("time,dim,dimLong,dimFloat,val\n");
        // unparseable
        writer.write("unparseable,a,2,3.0,1\n");
        // valid row
        writer.write("2014-01-01T00:00:10Z,a,2,3.0,1\n");
        // unparseable
        writer.write("9.0,a,2,3.0,1\n");
        // thrown away
        writer.write("3014-03-01T00:00:10Z,outsideofinterval,2,3.0,1\n");
        // unparseable
        writer.write("99999999999-01-01T00:00:10Z,b,2,3.0,1\n");
    }
    // Allow up to 3 parse exceptions, and save up to 2 parse exceptions
    final IndexTuningConfig tuningConfig = new IndexTuningConfig(null, null, null, null, null, null, null, null, null, null, new HashedPartitionsSpec(2, null, null), INDEX_SPEC, null, null, true, false, null, null, null, true, 2, 5, null, null);
    final TimestampSpec timestampSpec = new TimestampSpec("time", "auto", null);
    final DimensionsSpec dimensionsSpec = new DimensionsSpec(Arrays.asList(new StringDimensionSchema("dim"), new LongDimensionSchema("dimLong"), new FloatDimensionSchema("dimFloat")));
    final List<String> columns = Arrays.asList("time", "dim", "dimLong", "dimFloat", "val");
    final IndexIngestionSpec ingestionSpec;
    List<String> expectedMessages;
    if (useInputFormatApi) {
        ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, timestampSpec, dimensionsSpec, new CsvInputFormat(columns, null, null, true, 0), null, null, tuningConfig, false, false);
        expectedMessages = Arrays.asList(StringUtils.format("Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 3, Line: 6)", tmpFile.toURI()), StringUtils.format("Timestamp[9.0] is unparseable! Event: {time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 2, Line: 4)", tmpFile.toURI()), StringUtils.format("Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1} (Path: %s, Record: 1, Line: 2)", tmpFile.toURI()));
    } else {
        ingestionSpec = createIngestionSpec(jsonMapper, tmpDir, new CSVParseSpec(timestampSpec, dimensionsSpec, null, columns, true, 0), null, null, tuningConfig, false, false);
        expectedMessages = Arrays.asList("Timestamp[99999999999-01-01T00:00:10Z] is unparseable! Event: {time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "Timestamp[9.0] is unparseable! Event: {time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1}", "Timestamp[unparseable] is unparseable! Event: {time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
    }
    IndexTask indexTask = new IndexTask(null, null, ingestionSpec, null);
    TaskStatus status = runTask(indexTask).lhs;
    Assert.assertEquals(TaskState.FAILED, status.getStatusCode());
    checkTaskStatusErrorMsgForParseExceptionsExceeded(status);
    IngestionStatsAndErrorsTaskReportData reportData = getTaskReportData();
    Map<String, Object> expectedMetrics = ImmutableMap.of(RowIngestionMeters.DETERMINE_PARTITIONS, ImmutableMap.of(RowIngestionMeters.PROCESSED_WITH_ERROR, 0, RowIngestionMeters.PROCESSED, 1, RowIngestionMeters.UNPARSEABLE, 3, RowIngestionMeters.THROWN_AWAY, useInputFormatApi ? 1 : 2), RowIngestionMeters.BUILD_SEGMENTS, ImmutableMap.of(RowIngestionMeters.PROCESSED_WITH_ERROR, 0, RowIngestionMeters.PROCESSED, 0, RowIngestionMeters.UNPARSEABLE, 0, RowIngestionMeters.THROWN_AWAY, 0));
    Assert.assertEquals(expectedMetrics, reportData.getRowStats());
    List<LinkedHashMap> parseExceptionReports = (List<LinkedHashMap>) reportData.getUnparseableEvents().get(RowIngestionMeters.DETERMINE_PARTITIONS);
    List<String> actualMessages = parseExceptionReports.stream().map((r) -> {
        return ((List<String>) r.get("details")).get(0);
    }).collect(Collectors.toList());
    Assert.assertEquals(expectedMessages, actualMessages);
    List<String> expectedInputs = Arrays.asList("{time=99999999999-01-01T00:00:10Z, dim=b, dimLong=2, dimFloat=3.0, val=1}", "{time=9.0, dim=a, dimLong=2, dimFloat=3.0, val=1}", "{time=unparseable, dim=a, dimLong=2, dimFloat=3.0, val=1}");
    List<String> actualInputs = parseExceptionReports.stream().map((r) -> {
        return (String) r.get("input");
    }).collect(Collectors.toList());
    Assert.assertEquals(expectedInputs, actualInputs);
}

Example 19 with CsvInputFormat

use of org.apache.druid.data.input.impl.CsvInputFormat in project druid by druid-io.

the class IndexTaskTest method testIgnoreParseException.

@Test
public void testIgnoreParseException() throws Exception {
    final File tmpDir = temporaryFolder.newFolder();
    final File tmpFile = File.createTempFile("druid", "index", tmpDir);
    try (BufferedWriter writer = Files.newWriter(tmpFile, StandardCharsets.UTF_8)) {
        writer.write("time,d,val\n");
        writer.write("unparseable,a,1\n");
        writer.write("2014-01-01T00:00:10Z,a,1\n");
    }
    final TimestampSpec timestampSpec = new TimestampSpec("time", "auto", null);
    final List<String> columns = Arrays.asList("time", "dim", "val");
    // ignore parse exception
    final IndexTuningConfig tuningConfig = createTuningConfig(2, null, null, null, null, false, false);
    // GranularitySpec.intervals and numShards must be null to verify reportParseException=false is respected both in
    // IndexTask.determineShardSpecs() and IndexTask.generateAndPublishSegments()
    final IndexIngestionSpec parseExceptionIgnoreSpec;
    if (useInputFormatApi) {
        parseExceptionIgnoreSpec = createIngestionSpec(jsonMapper, tmpDir, timestampSpec, DimensionsSpec.EMPTY, new CsvInputFormat(columns, null, null, true, 0), null, null, tuningConfig, false, false);
    } else {
        parseExceptionIgnoreSpec = createIngestionSpec(jsonMapper, tmpDir, new CSVParseSpec(timestampSpec, DimensionsSpec.EMPTY, null, columns, true, 0), null, null, tuningConfig, false, false);
    }
    IndexTask indexTask = new IndexTask(null, null, parseExceptionIgnoreSpec, null);
    final List<DataSegment> segments = runTask(indexTask).rhs;
    Assert.assertEquals(Collections.singletonList("d"), segments.get(0).getDimensions());
    Assert.assertEquals(Collections.singletonList("val"), segments.get(0).getMetrics());
    Assert.assertEquals(Intervals.of("2014/P1D"), segments.get(0).getInterval());
}