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

Example 56 with InputRowSchema

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

the class S3InputSourceTest method testCompressedReader.

@Test
public void testCompressedReader() throws IOException {
    EasyMock.reset(S3_CLIENT);
    expectListObjects(PREFIXES.get(0), ImmutableList.of(EXPECTED_COMPRESSED_URIS.get(0)), CONTENT);
    expectListObjects(EXPECTED_COMPRESSED_URIS.get(1), ImmutableList.of(EXPECTED_COMPRESSED_URIS.get(1)), CONTENT);
    expectGetObjectCompressed(EXPECTED_COMPRESSED_URIS.get(0));
    expectGetObjectCompressed(EXPECTED_COMPRESSED_URIS.get(1));
    EasyMock.replay(S3_CLIENT);
    S3InputSource inputSource = new S3InputSource(SERVICE, SERVER_SIDE_ENCRYPTING_AMAZON_S3_BUILDER, INPUT_DATA_CONFIG, null, ImmutableList.of(PREFIXES.get(0), EXPECTED_COMPRESSED_URIS.get(1)), null, null);
    InputRowSchema someSchema = new InputRowSchema(new TimestampSpec("time", "auto", null), new DimensionsSpec(DimensionsSpec.getDefaultSchemas(ImmutableList.of("dim1", "dim2"))), ColumnsFilter.all());
    InputSourceReader reader = inputSource.reader(someSchema, new CsvInputFormat(ImmutableList.of("time", "dim1", "dim2"), "|", false, null, 0), temporaryFolder.newFolder());
    CloseableIterator<InputRow> iterator = reader.read();
    while (iterator.hasNext()) {
        InputRow nextRow = iterator.next();
        Assert.assertEquals(NOW, nextRow.getTimestamp());
        Assert.assertEquals("hello", nextRow.getDimension("dim1").get(0));
        Assert.assertEquals("world", nextRow.getDimension("dim2").get(0));
    }
    EasyMock.verify(S3_CLIENT);
}

Example 57 with InputRowSchema

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

the class DruidInputSource method fixedFormatReader.

@Override
protected InputSourceReader fixedFormatReader(InputRowSchema inputRowSchema, @Nullable File temporaryDirectory) {
    final SegmentCacheManager segmentCacheManager = segmentCacheManagerFactory.manufacturate(temporaryDirectory);
    final List<TimelineObjectHolder<String, DataSegment>> timeline = createTimeline();
    final Iterator<DruidSegmentInputEntity> entityIterator = FluentIterable.from(timeline).transformAndConcat(holder -> {
        // noinspection ConstantConditions
        final PartitionHolder<DataSegment> partitionHolder = holder.getObject();
        // noinspection ConstantConditions
        return FluentIterable.from(partitionHolder).transform(chunk -> new DruidSegmentInputEntity(segmentCacheManager, chunk.getObject(), holder.getInterval()));
    }).iterator();
    final DruidSegmentInputFormat inputFormat = new DruidSegmentInputFormat(indexIO, dimFilter);
    final InputRowSchema inputRowSchemaToUse;
    if (taskConfig.isIgnoreTimestampSpecForDruidInputSource()) {
        // Legacy compatibility mode; see https://github.com/apache/druid/pull/10267.
        LOG.warn("Ignoring the provided timestampSpec and reading the __time column instead. To use timestampSpecs with " + "the 'druid' input source, set druid.indexer.task.ignoreTimestampSpecForDruidInputSource to false.");
        inputRowSchemaToUse = new InputRowSchema(new TimestampSpec(ColumnHolder.TIME_COLUMN_NAME, STANDARD_TIME_COLUMN_FORMATS.iterator().next(), null), inputRowSchema.getDimensionsSpec(), inputRowSchema.getColumnsFilter().plus(ColumnHolder.TIME_COLUMN_NAME));
    } else {
        inputRowSchemaToUse = inputRowSchema;
    }
    if (ColumnHolder.TIME_COLUMN_NAME.equals(inputRowSchemaToUse.getTimestampSpec().getTimestampColumn()) && !STANDARD_TIME_COLUMN_FORMATS.contains(inputRowSchemaToUse.getTimestampSpec().getTimestampFormat())) {
        // Slight chance the user did this intentionally, but not likely. Log a warning.
        LOG.warn("The provided timestampSpec refers to the %s column without using format %s. If you wanted to read the " + "column as-is, switch formats.", inputRowSchemaToUse.getTimestampSpec().getTimestampColumn(), STANDARD_TIME_COLUMN_FORMATS);
    }
    return new InputEntityIteratingReader(inputRowSchemaToUse, inputFormat, entityIterator, temporaryDirectory);
}

Example 58 with InputRowSchema

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

the class KafkaInputFormatTest method testWithHeaderKeyAndValue.

@Test
public void testWithHeaderKeyAndValue() throws IOException {
    final byte[] key = StringUtils.toUtf8("{\n" + "    \"key\": \"sampleKey\"\n" + "}");
    final byte[] payload = StringUtils.toUtf8("{\n" + "    \"timestamp\": \"2021-06-25\",\n" + "    \"bar\": null,\n" + "    \"foo\": \"x\",\n" + "    \"baz\": 4,\n" + "    \"o\": {\n" + "        \"mg\": 1\n" + "    }\n" + "}");
    Headers headers = new RecordHeaders(SAMPLE_HEADERS);
    inputEntity = new KafkaRecordEntity(new ConsumerRecord<byte[], byte[]>("sample", 0, 0, timestamp, null, null, 0, 0, key, payload, headers));
    final InputEntityReader reader = format.createReader(new InputRowSchema(new TimestampSpec("timestamp", "iso", null), new DimensionsSpec(DimensionsSpec.getDefaultSchemas(ImmutableList.of("bar", "foo", "kafka.newheader.encoding", "kafka.newheader.kafkapkc", "kafka.newts.timestamp"))), ColumnsFilter.all()), inputEntity, null);
    final int numExpectedIterations = 1;
    try (CloseableIterator<InputRow> iterator = reader.read()) {
        int numActualIterations = 0;
        while (iterator.hasNext()) {
            final InputRow row = iterator.next();
            // Payload verifications
            Assert.assertEquals(DateTimes.of("2021-06-25"), row.getTimestamp());
            Assert.assertEquals("x", Iterables.getOnlyElement(row.getDimension("foo")));
            Assert.assertEquals("4", Iterables.getOnlyElement(row.getDimension("baz")));
            Assert.assertEquals("4", Iterables.getOnlyElement(row.getDimension("root_baz")));
            Assert.assertEquals("1", Iterables.getOnlyElement(row.getDimension("path_omg")));
            Assert.assertEquals("1", Iterables.getOnlyElement(row.getDimension("jq_omg")));
            // Header verification
            Assert.assertEquals("application/json", Iterables.getOnlyElement(row.getDimension("kafka.newheader.encoding")));
            Assert.assertEquals("pkc-bar", Iterables.getOnlyElement(row.getDimension("kafka.newheader.kafkapkc")));
            Assert.assertEquals(String.valueOf(DateTimes.of("2021-06-24").getMillis()), Iterables.getOnlyElement(row.getDimension("kafka.newts.timestamp")));
            Assert.assertEquals("2021-06-25", Iterables.getOnlyElement(row.getDimension("timestamp")));
            // Key verification
            Assert.assertEquals("sampleKey", Iterables.getOnlyElement(row.getDimension("kafka.newkey.key")));
            Assert.assertTrue(row.getDimension("root_baz2").isEmpty());
            Assert.assertTrue(row.getDimension("path_omg2").isEmpty());
            Assert.assertTrue(row.getDimension("jq_omg2").isEmpty());
            numActualIterations++;
        }
        Assert.assertEquals(numExpectedIterations, numActualIterations);
    }
}

Example 59 with InputRowSchema

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

the class KafkaInputFormatTest method testWithOutKeyAndHeaderSpecs.

@Test
public void testWithOutKeyAndHeaderSpecs() throws IOException {
    final byte[] payload = StringUtils.toUtf8("{\n" + "    \"timestamp\": \"2021-06-24\",\n" + "    \"bar\": null,\n" + "    \"foo\": \"x\",\n" + "    \"baz\": 4,\n" + "    \"o\": {\n" + "        \"mg\": 1\n" + "    }\n" + "}");
    Headers headers = new RecordHeaders(SAMPLE_HEADERS);
    inputEntity = new KafkaRecordEntity(new ConsumerRecord<byte[], byte[]>("sample", 0, 0, timestamp, null, null, 0, 0, null, payload, headers));
    KafkaInputFormat localFormat = new KafkaInputFormat(null, null, // Value Format
    new JsonInputFormat(new JSONPathSpec(true, ImmutableList.of(new JSONPathFieldSpec(JSONPathFieldType.ROOT, "root_baz", "baz"), new JSONPathFieldSpec(JSONPathFieldType.ROOT, "root_baz2", "baz2"), new JSONPathFieldSpec(JSONPathFieldType.PATH, "path_omg", "$.o.mg"), new JSONPathFieldSpec(JSONPathFieldType.PATH, "path_omg2", "$.o.mg2"), new JSONPathFieldSpec(JSONPathFieldType.JQ, "jq_omg", ".o.mg"), new JSONPathFieldSpec(JSONPathFieldType.JQ, "jq_omg2", ".o.mg2"))), // make sure JsonReader is used
    null, // make sure JsonReader is used
    null, // make sure JsonReader is used
    false), "kafka.newheader.", "kafka.newkey.", "kafka.newts.");
    final InputEntityReader reader = localFormat.createReader(new InputRowSchema(new TimestampSpec("timestamp", "iso", null), new DimensionsSpec(DimensionsSpec.getDefaultSchemas(ImmutableList.of("bar", "foo", "kafka.newts.timestamp"))), ColumnsFilter.all()), inputEntity, null);
    final int numExpectedIterations = 1;
    try (CloseableIterator<InputRow> iterator = reader.read()) {
        int numActualIterations = 0;
        while (iterator.hasNext()) {
            final InputRow row = iterator.next();
            // Key verification
            Assert.assertTrue(row.getDimension("kafka.newkey.key").isEmpty());
            Assert.assertEquals("x", Iterables.getOnlyElement(row.getDimension("foo")));
            Assert.assertEquals("4", Iterables.getOnlyElement(row.getDimension("baz")));
            Assert.assertEquals("4", Iterables.getOnlyElement(row.getDimension("root_baz")));
            Assert.assertEquals("1", Iterables.getOnlyElement(row.getDimension("path_omg")));
            Assert.assertEquals("1", Iterables.getOnlyElement(row.getDimension("jq_omg")));
            numActualIterations++;
        }
        Assert.assertEquals(numExpectedIterations, numActualIterations);
    }
}

Example 60 with InputRowSchema

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

the class KafkaInputFormatTest method testWithOutKey.

@Test
public // Headers cannot be null, so testing only no key use case!
void testWithOutKey() throws IOException {
    final byte[] payload = StringUtils.toUtf8("{\n" + "    \"timestamp\": \"2021-06-24\",\n" + "    \"bar\": null,\n" + "    \"foo\": \"x\",\n" + "    \"baz\": 4,\n" + "    \"o\": {\n" + "        \"mg\": 1\n" + "    }\n" + "}");
    Headers headers = new RecordHeaders(SAMPLE_HEADERS);
    inputEntity = new KafkaRecordEntity(new ConsumerRecord<byte[], byte[]>("sample", 0, 0, timestamp, null, null, 0, 0, null, payload, headers));
    final InputEntityReader reader = format.createReader(new InputRowSchema(new TimestampSpec("timestamp", "iso", null), new DimensionsSpec(DimensionsSpec.getDefaultSchemas(ImmutableList.of("bar", "foo", "kafka.newheader.encoding", "kafka.newheader.kafkapkc", "kafka.newts.timestamp"))), ColumnsFilter.all()), inputEntity, null);
    final int numExpectedIterations = 1;
    try (CloseableIterator<InputRow> iterator = reader.read()) {
        int numActualIterations = 0;
        while (iterator.hasNext()) {
            final InputRow row = iterator.next();
            // Key verification
            Assert.assertTrue(row.getDimension("kafka.newkey.key").isEmpty());
            numActualIterations++;
        }
        Assert.assertEquals(numExpectedIterations, numActualIterations);
    }
}