Examples with RecordSetWriter org.apache.nifi.serialization.RecordSetWriter used on opensource projects

Example 6 with RecordSetWriter

use of org.apache.nifi.serialization.RecordSetWriter in project nifi by apache.

the class FetchParquetTest method testIOExceptionWhileWritingShouldRouteToRetry.

@Test
public void testIOExceptionWhileWritingShouldRouteToRetry() throws InitializationException, IOException, SchemaNotFoundException {
    configure(proc);
    final RecordSetWriter recordSetWriter = Mockito.mock(RecordSetWriter.class);
    when(recordSetWriter.write(any(Record.class))).thenThrow(new IOException("IOException"));
    final RecordSetWriterFactory recordSetWriterFactory = Mockito.mock(RecordSetWriterFactory.class);
    when(recordSetWriterFactory.getIdentifier()).thenReturn("mock-writer-factory");
    when(recordSetWriterFactory.createWriter(any(ComponentLog.class), any(RecordSchema.class), any(OutputStream.class))).thenReturn(recordSetWriter);
    testRunner.addControllerService("mock-writer-factory", recordSetWriterFactory);
    testRunner.enableControllerService(recordSetWriterFactory);
    testRunner.setProperty(FetchParquet.RECORD_WRITER, "mock-writer-factory");
    final File parquetDir = new File(DIRECTORY);
    final File parquetFile = new File(parquetDir, "testFetchParquetToCSV.parquet");
    final int numUsers = 10;
    writeParquetUsers(parquetFile, numUsers);
    final Map<String, String> attributes = new HashMap<>();
    attributes.put(CoreAttributes.PATH.key(), parquetDir.getAbsolutePath());
    attributes.put(CoreAttributes.FILENAME.key(), parquetFile.getName());
    testRunner.enqueue("TRIGGER", attributes);
    testRunner.run();
    testRunner.assertAllFlowFilesTransferred(FetchParquet.REL_RETRY, 1);
    final MockFlowFile flowFile = testRunner.getFlowFilesForRelationship(FetchParquet.REL_RETRY).get(0);
    flowFile.assertContentEquals("TRIGGER");
}

Example 7 with RecordSetWriter

use of org.apache.nifi.serialization.RecordSetWriter in project nifi by apache.

the class SplitRecord method onTrigger.

@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
    FlowFile original = session.get();
    if (original == null) {
        return;
    }
    final RecordReaderFactory readerFactory = context.getProperty(RECORD_READER).asControllerService(RecordReaderFactory.class);
    final RecordSetWriterFactory writerFactory = context.getProperty(RECORD_WRITER).asControllerService(RecordSetWriterFactory.class);
    final int maxRecords = context.getProperty(RECORDS_PER_SPLIT).evaluateAttributeExpressions(original).asInteger();
    final List<FlowFile> splits = new ArrayList<>();
    final Map<String, String> originalAttributes = original.getAttributes();
    try {
        session.read(original, new InputStreamCallback() {

            @Override
            public void process(final InputStream in) throws IOException {
                try (final RecordReader reader = readerFactory.createRecordReader(originalAttributes, in, getLogger())) {
                    final RecordSchema schema = writerFactory.getSchema(originalAttributes, reader.getSchema());
                    final RecordSet recordSet = reader.createRecordSet();
                    final PushBackRecordSet pushbackSet = new PushBackRecordSet(recordSet);
                    while (pushbackSet.isAnotherRecord()) {
                        FlowFile split = session.create(original);
                        try {
                            final Map<String, String> attributes = new HashMap<>();
                            final WriteResult writeResult;
                            try (final OutputStream out = session.write(split);
                                final RecordSetWriter writer = writerFactory.createWriter(getLogger(), schema, out)) {
                                if (maxRecords == 1) {
                                    final Record record = pushbackSet.next();
                                    writeResult = writer.write(record);
                                } else {
                                    final RecordSet limitedSet = pushbackSet.limit(maxRecords);
                                    writeResult = writer.write(limitedSet);
                                }
                                attributes.put("record.count", String.valueOf(writeResult.getRecordCount()));
                                attributes.put(CoreAttributes.MIME_TYPE.key(), writer.getMimeType());
                                attributes.putAll(writeResult.getAttributes());
                                session.adjustCounter("Records Split", writeResult.getRecordCount(), false);
                            }
                            split = session.putAllAttributes(split, attributes);
                        } finally {
                            splits.add(split);
                        }
                    }
                } catch (final SchemaNotFoundException | MalformedRecordException e) {
                    throw new ProcessException("Failed to parse incoming data", e);
                }
            }
        });
    } catch (final ProcessException pe) {
        getLogger().error("Failed to split {}", new Object[] { original, pe });
        session.remove(splits);
        session.transfer(original, REL_FAILURE);
        return;
    }
    session.transfer(original, REL_ORIGINAL);
    session.transfer(splits, REL_SPLITS);
    getLogger().info("Successfully split {} into {} FlowFiles, each containing up to {} records", new Object[] { original, splits.size(), maxRecords });
}

Example 8 with RecordSetWriter

use of org.apache.nifi.serialization.RecordSetWriter in project nifi by apache.

the class ValidateRecord method onTrigger.

@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
    FlowFile flowFile = session.get();
    if (flowFile == null) {
        return;
    }
    final RecordSetWriterFactory writerFactory = context.getProperty(RECORD_WRITER).asControllerService(RecordSetWriterFactory.class);
    final RecordReaderFactory readerFactory = context.getProperty(RECORD_READER).asControllerService(RecordReaderFactory.class);
    final boolean allowExtraFields = context.getProperty(ALLOW_EXTRA_FIELDS).asBoolean();
    final boolean strictTypeChecking = context.getProperty(STRICT_TYPE_CHECKING).asBoolean();
    RecordSetWriter validWriter = null;
    RecordSetWriter invalidWriter = null;
    FlowFile validFlowFile = null;
    FlowFile invalidFlowFile = null;
    try (final InputStream in = session.read(flowFile);
        final RecordReader reader = readerFactory.createRecordReader(flowFile, in, getLogger())) {
        final RecordSchema validationSchema = getValidationSchema(context, flowFile, reader);
        final SchemaValidationContext validationContext = new SchemaValidationContext(validationSchema, allowExtraFields, strictTypeChecking);
        final RecordSchemaValidator validator = new StandardSchemaValidator(validationContext);
        int recordCount = 0;
        int validCount = 0;
        int invalidCount = 0;
        final Set<String> extraFields = new HashSet<>();
        final Set<String> missingFields = new HashSet<>();
        final Set<String> invalidFields = new HashSet<>();
        final Set<String> otherProblems = new HashSet<>();
        try {
            Record record;
            while ((record = reader.nextRecord(false, false)) != null) {
                final SchemaValidationResult result = validator.validate(record);
                recordCount++;
                RecordSetWriter writer;
                if (result.isValid()) {
                    validCount++;
                    if (validFlowFile == null) {
                        validFlowFile = session.create(flowFile);
                    }
                    validWriter = writer = createIfNecessary(validWriter, writerFactory, session, validFlowFile, record.getSchema());
                } else {
                    invalidCount++;
                    logValidationErrors(flowFile, recordCount, result);
                    if (invalidFlowFile == null) {
                        invalidFlowFile = session.create(flowFile);
                    }
                    invalidWriter = writer = createIfNecessary(invalidWriter, writerFactory, session, invalidFlowFile, record.getSchema());
                    // that it is too noisy to be useful.
                    for (final ValidationError validationError : result.getValidationErrors()) {
                        final Optional<String> fieldName = validationError.getFieldName();
                        switch(validationError.getType()) {
                            case EXTRA_FIELD:
                                if (fieldName.isPresent()) {
                                    extraFields.add(fieldName.get());
                                } else {
                                    otherProblems.add(validationError.getExplanation());
                                }
                                break;
                            case MISSING_FIELD:
                                if (fieldName.isPresent()) {
                                    missingFields.add(fieldName.get());
                                } else {
                                    otherProblems.add(validationError.getExplanation());
                                }
                                break;
                            case INVALID_FIELD:
                                if (fieldName.isPresent()) {
                                    invalidFields.add(fieldName.get());
                                } else {
                                    otherProblems.add(validationError.getExplanation());
                                }
                                break;
                            case OTHER:
                                otherProblems.add(validationError.getExplanation());
                                break;
                        }
                    }
                }
                if (writer instanceof RawRecordWriter) {
                    ((RawRecordWriter) writer).writeRawRecord(record);
                } else {
                    writer.write(record);
                }
            }
            if (validWriter != null) {
                completeFlowFile(session, validFlowFile, validWriter, REL_VALID, null);
            }
            if (invalidWriter != null) {
                // Build up a String that explains why the records were invalid, so that we can add this to the Provenance Event.
                final StringBuilder errorBuilder = new StringBuilder();
                errorBuilder.append("Records in this FlowFile were invalid for the following reasons: ");
                if (!missingFields.isEmpty()) {
                    errorBuilder.append("The following ").append(missingFields.size()).append(" fields were missing: ").append(missingFields.toString());
                }
                if (!extraFields.isEmpty()) {
                    if (errorBuilder.length() > 0) {
                        errorBuilder.append("; ");
                    }
                    errorBuilder.append("The following ").append(extraFields.size()).append(" fields were present in the Record but not in the schema: ").append(extraFields.toString());
                }
                if (!invalidFields.isEmpty()) {
                    if (errorBuilder.length() > 0) {
                        errorBuilder.append("; ");
                    }
                    errorBuilder.append("The following ").append(invalidFields.size()).append(" fields had values whose type did not match the schema: ").append(invalidFields.toString());
                }
                if (!otherProblems.isEmpty()) {
                    if (errorBuilder.length() > 0) {
                        errorBuilder.append("; ");
                    }
                    errorBuilder.append("The following ").append(otherProblems.size()).append(" additional problems were encountered: ").append(otherProblems.toString());
                }
                final String validationErrorString = errorBuilder.toString();
                completeFlowFile(session, invalidFlowFile, invalidWriter, REL_INVALID, validationErrorString);
            }
        } finally {
            closeQuietly(validWriter);
            closeQuietly(invalidWriter);
        }
        session.adjustCounter("Records Validated", recordCount, false);
        session.adjustCounter("Records Found Valid", validCount, false);
        session.adjustCounter("Records Found Invalid", invalidCount, false);
    } catch (final IOException | MalformedRecordException | SchemaNotFoundException e) {
        getLogger().error("Failed to process {}; will route to failure", new Object[] { flowFile, e });
        session.transfer(flowFile, REL_FAILURE);
        if (validFlowFile != null) {
            session.remove(validFlowFile);
        }
        if (invalidFlowFile != null) {
            session.remove(invalidFlowFile);
        }
        return;
    }
    session.remove(flowFile);
}

Example 9 with RecordSetWriter

use of org.apache.nifi.serialization.RecordSetWriter in project nifi by apache.

the class AbstractRecordProcessor method onTrigger.

@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
    FlowFile flowFile = session.get();
    if (flowFile == null) {
        return;
    }
    final RecordReaderFactory readerFactory = context.getProperty(RECORD_READER).asControllerService(RecordReaderFactory.class);
    final RecordSetWriterFactory writerFactory = context.getProperty(RECORD_WRITER).asControllerService(RecordSetWriterFactory.class);
    final Map<String, String> attributes = new HashMap<>();
    final AtomicInteger recordCount = new AtomicInteger();
    final FlowFile original = flowFile;
    final Map<String, String> originalAttributes = flowFile.getAttributes();
    try {
        flowFile = session.write(flowFile, new StreamCallback() {

            @Override
            public void process(final InputStream in, final OutputStream out) throws IOException {
                try (final RecordReader reader = readerFactory.createRecordReader(originalAttributes, in, getLogger())) {
                    final RecordSchema writeSchema = writerFactory.getSchema(originalAttributes, reader.getSchema());
                    try (final RecordSetWriter writer = writerFactory.createWriter(getLogger(), writeSchema, out)) {
                        writer.beginRecordSet();
                        Record record;
                        while ((record = reader.nextRecord()) != null) {
                            final Record processed = AbstractRecordProcessor.this.process(record, writeSchema, original, context);
                            writer.write(processed);
                        }
                        final WriteResult writeResult = writer.finishRecordSet();
                        attributes.put("record.count", String.valueOf(writeResult.getRecordCount()));
                        attributes.put(CoreAttributes.MIME_TYPE.key(), writer.getMimeType());
                        attributes.putAll(writeResult.getAttributes());
                        recordCount.set(writeResult.getRecordCount());
                    }
                } catch (final SchemaNotFoundException e) {
                    throw new ProcessException(e.getLocalizedMessage(), e);
                } catch (final MalformedRecordException e) {
                    throw new ProcessException("Could not parse incoming data", e);
                }
            }
        });
    } catch (final Exception e) {
        getLogger().error("Failed to process {}; will route to failure", new Object[] { flowFile, e });
        session.transfer(flowFile, REL_FAILURE);
        return;
    }
    flowFile = session.putAllAttributes(flowFile, attributes);
    session.transfer(flowFile, REL_SUCCESS);
    final int count = recordCount.get();
    session.adjustCounter("Records Processed", count, false);
    getLogger().info("Successfully converted {} records for {}", new Object[] { count, flowFile });
}

Example 10 with RecordSetWriter

use of org.apache.nifi.serialization.RecordSetWriter in project nifi by apache.

the class AbstractRouteRecord method onTrigger.

@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
    FlowFile flowFile = session.get();
    if (flowFile == null) {
        return;
    }
    final T flowFileContext;
    try {
        flowFileContext = getFlowFileContext(flowFile, context);
    } catch (final Exception e) {
        getLogger().error("Failed to process {}; routing to failure", new Object[] { flowFile, e });
        session.transfer(flowFile, REL_FAILURE);
        return;
    }
    final RecordReaderFactory readerFactory = context.getProperty(RECORD_READER).asControllerService(RecordReaderFactory.class);
    final RecordSetWriterFactory writerFactory = context.getProperty(RECORD_WRITER).asControllerService(RecordSetWriterFactory.class);
    final AtomicInteger numRecords = new AtomicInteger(0);
    final Map<Relationship, Tuple<FlowFile, RecordSetWriter>> writers = new HashMap<>();
    final FlowFile original = flowFile;
    final Map<String, String> originalAttributes = original.getAttributes();
    try {
        session.read(flowFile, new InputStreamCallback() {

            @Override
            public void process(final InputStream in) throws IOException {
                try (final RecordReader reader = readerFactory.createRecordReader(originalAttributes, in, getLogger())) {
                    final RecordSchema writeSchema = writerFactory.getSchema(originalAttributes, reader.getSchema());
                    Record record;
                    while ((record = reader.nextRecord()) != null) {
                        final Set<Relationship> relationships = route(record, writeSchema, original, context, flowFileContext);
                        numRecords.incrementAndGet();
                        for (final Relationship relationship : relationships) {
                            final RecordSetWriter recordSetWriter;
                            Tuple<FlowFile, RecordSetWriter> tuple = writers.get(relationship);
                            if (tuple == null) {
                                FlowFile outFlowFile = session.create(original);
                                final OutputStream out = session.write(outFlowFile);
                                recordSetWriter = writerFactory.createWriter(getLogger(), writeSchema, out);
                                recordSetWriter.beginRecordSet();
                                tuple = new Tuple<>(outFlowFile, recordSetWriter);
                                writers.put(relationship, tuple);
                            } else {
                                recordSetWriter = tuple.getValue();
                            }
                            recordSetWriter.write(record);
                        }
                    }
                } catch (final SchemaNotFoundException | MalformedRecordException e) {
                    throw new ProcessException("Could not parse incoming data", e);
                }
            }
        });
        for (final Map.Entry<Relationship, Tuple<FlowFile, RecordSetWriter>> entry : writers.entrySet()) {
            final Relationship relationship = entry.getKey();
            final Tuple<FlowFile, RecordSetWriter> tuple = entry.getValue();
            final RecordSetWriter writer = tuple.getValue();
            FlowFile childFlowFile = tuple.getKey();
            final WriteResult writeResult = writer.finishRecordSet();
            try {
                writer.close();
            } catch (final IOException ioe) {
                getLogger().warn("Failed to close Writer for {}", new Object[] { childFlowFile });
            }
            final Map<String, String> attributes = new HashMap<>();
            attributes.put("record.count", String.valueOf(writeResult.getRecordCount()));
            attributes.put(CoreAttributes.MIME_TYPE.key(), writer.getMimeType());
            attributes.putAll(writeResult.getAttributes());
            childFlowFile = session.putAllAttributes(childFlowFile, attributes);
            session.transfer(childFlowFile, relationship);
            session.adjustCounter("Records Processed", writeResult.getRecordCount(), false);
            session.adjustCounter("Records Routed to " + relationship.getName(), writeResult.getRecordCount(), false);
            session.getProvenanceReporter().route(childFlowFile, relationship);
        }
    } catch (final Exception e) {
        getLogger().error("Failed to process {}", new Object[] { flowFile, e });
        for (final Tuple<FlowFile, RecordSetWriter> tuple : writers.values()) {
            try {
                tuple.getValue().close();
            } catch (final Exception e1) {
                getLogger().warn("Failed to close Writer for {}; some resources may not be cleaned up appropriately", new Object[] { tuple.getKey() });
            }
            session.remove(tuple.getKey());
        }
        session.transfer(flowFile, REL_FAILURE);
        return;
    } finally {
        for (final Tuple<FlowFile, RecordSetWriter> tuple : writers.values()) {
            final RecordSetWriter writer = tuple.getValue();
            try {
                writer.close();
            } catch (final Exception e) {
                getLogger().warn("Failed to close Record Writer for {}; some resources may not be properly cleaned up", new Object[] { tuple.getKey(), e });
            }
        }
    }
    if (isRouteOriginal()) {
        flowFile = session.putAttribute(flowFile, "record.count", String.valueOf(numRecords));
        session.transfer(flowFile, REL_ORIGINAL);
    } else {
        session.remove(flowFile);
    }
    getLogger().info("Successfully processed {}, creating {} derivative FlowFiles and processing {} records", new Object[] { flowFile, writers.size(), numRecords });
}