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

Example 11 with RecordSetWriterFactory

use of org.apache.nifi.serialization.RecordSetWriterFactory 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 12 with RecordSetWriterFactory

use of org.apache.nifi.serialization.RecordSetWriterFactory 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 13 with RecordSetWriterFactory

use of org.apache.nifi.serialization.RecordSetWriterFactory 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 });
}

Example 14 with RecordSetWriterFactory

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

the class GetSolr method onTrigger.

@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
    final ComponentLog logger = getLogger();
    final AtomicBoolean continuePaging = new AtomicBoolean(true);
    final SolrQuery solrQuery = new SolrQuery();
    try {
        if (id_field == null) {
            id_field = getFieldNameOfUniqueKey();
        }
        final String dateField = context.getProperty(DATE_FIELD).getValue();
        final Map<String, String> stateMap = new HashMap<String, String>();
        stateMap.putAll(context.getStateManager().getState(Scope.CLUSTER).toMap());
        solrQuery.setQuery("*:*");
        final String query = context.getProperty(SOLR_QUERY).getValue();
        if (!StringUtils.isBlank(query) && !query.equals("*:*")) {
            solrQuery.addFilterQuery(query);
        }
        final StringBuilder automatedFilterQuery = (new StringBuilder()).append(dateField).append(":[").append(stateMap.get(STATE_MANAGER_FILTER)).append(" TO *]");
        solrQuery.addFilterQuery(automatedFilterQuery.toString());
        final List<String> fieldList = new ArrayList<String>();
        final String returnFields = context.getProperty(RETURN_FIELDS).getValue();
        if (!StringUtils.isBlank(returnFields)) {
            fieldList.addAll(Arrays.asList(returnFields.trim().split("[,]")));
            if (!fieldList.contains(dateField)) {
                fieldList.add(dateField);
                dateFieldNotInSpecifiedFieldsList.set(true);
            }
            for (String returnField : fieldList) {
                solrQuery.addField(returnField.trim());
            }
        }
        solrQuery.setParam(CursorMarkParams.CURSOR_MARK_PARAM, stateMap.get(STATE_MANAGER_CURSOR_MARK));
        solrQuery.setRows(context.getProperty(BATCH_SIZE).asInteger());
        final StringBuilder sortClause = (new StringBuilder()).append(dateField).append(" asc,").append(id_field).append(" asc");
        solrQuery.setParam("sort", sortClause.toString());
        while (continuePaging.get()) {
            final QueryRequest req = new QueryRequest(solrQuery);
            if (isBasicAuthEnabled()) {
                req.setBasicAuthCredentials(getUsername(), getPassword());
            }
            logger.debug(solrQuery.toQueryString());
            final QueryResponse response = req.process(getSolrClient());
            final SolrDocumentList documentList = response.getResults();
            if (response.getResults().size() > 0) {
                final SolrDocument lastSolrDocument = documentList.get(response.getResults().size() - 1);
                final String latestDateValue = df.format(lastSolrDocument.get(dateField));
                final String newCursorMark = response.getNextCursorMark();
                solrQuery.setParam(CursorMarkParams.CURSOR_MARK_PARAM, newCursorMark);
                stateMap.put(STATE_MANAGER_CURSOR_MARK, newCursorMark);
                stateMap.put(STATE_MANAGER_FILTER, latestDateValue);
                FlowFile flowFile = session.create();
                flowFile = session.putAttribute(flowFile, "solrQuery", solrQuery.toString());
                if (context.getProperty(RETURN_TYPE).getValue().equals(MODE_XML.getValue())) {
                    if (dateFieldNotInSpecifiedFieldsList.get()) {
                        for (SolrDocument doc : response.getResults()) {
                            doc.removeFields(dateField);
                        }
                    }
                    flowFile = session.write(flowFile, SolrUtils.getOutputStreamCallbackToTransformSolrResponseToXml(response));
                    flowFile = session.putAttribute(flowFile, CoreAttributes.MIME_TYPE.key(), "application/xml");
                } else {
                    final RecordSetWriterFactory writerFactory = context.getProperty(RECORD_WRITER).asControllerService(RecordSetWriterFactory.class);
                    final RecordSchema schema = writerFactory.getSchema(null, null);
                    final RecordSet recordSet = SolrUtils.solrDocumentsToRecordSet(response.getResults(), schema);
                    final StringBuffer mimeType = new StringBuffer();
                    flowFile = session.write(flowFile, new OutputStreamCallback() {

                        @Override
                        public void process(final OutputStream out) throws IOException {
                            try {
                                final RecordSetWriter writer = writerFactory.createWriter(getLogger(), schema, out);
                                writer.write(recordSet);
                                writer.flush();
                                mimeType.append(writer.getMimeType());
                            } catch (SchemaNotFoundException e) {
                                throw new ProcessException("Could not parse Solr response", e);
                            }
                        }
                    });
                    flowFile = session.putAttribute(flowFile, CoreAttributes.MIME_TYPE.key(), mimeType.toString());
                }
                session.transfer(flowFile, REL_SUCCESS);
            }
            continuePaging.set(response.getResults().size() == Integer.parseInt(context.getProperty(BATCH_SIZE).getValue()));
        }
        context.getStateManager().setState(stateMap, Scope.CLUSTER);
    } catch (SolrServerException | SchemaNotFoundException | IOException e) {
        context.yield();
        session.rollback();
        logger.error("Failed to execute query {} due to {}", new Object[] { solrQuery.toString(), e }, e);
        throw new ProcessException(e);
    } catch (final Throwable t) {
        context.yield();
        session.rollback();
        logger.error("Failed to execute query {} due to {}", new Object[] { solrQuery.toString(), t }, t);
        throw t;
    }
}

Example 15 with RecordSetWriterFactory

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

the class PublishKafkaRecord_0_10 method onTrigger.

@Override
public void onTrigger(final ProcessContext context, final ProcessSession session) throws ProcessException {
    final List<FlowFile> flowFiles = session.get(FlowFileFilters.newSizeBasedFilter(1, DataUnit.MB, 500));
    if (flowFiles.isEmpty()) {
        return;
    }
    final PublisherPool pool = getPublisherPool(context);
    if (pool == null) {
        context.yield();
        return;
    }
    final String securityProtocol = context.getProperty(KafkaProcessorUtils.SECURITY_PROTOCOL).getValue();
    final String bootstrapServers = context.getProperty(KafkaProcessorUtils.BOOTSTRAP_SERVERS).evaluateAttributeExpressions().getValue();
    final RecordSetWriterFactory writerFactory = context.getProperty(RECORD_WRITER).asControllerService(RecordSetWriterFactory.class);
    final RecordReaderFactory readerFactory = context.getProperty(RECORD_READER).asControllerService(RecordReaderFactory.class);
    final long startTime = System.nanoTime();
    try (final PublisherLease lease = pool.obtainPublisher()) {
        // Send each FlowFile to Kafka asynchronously.
        for (final FlowFile flowFile : flowFiles) {
            if (!isScheduled()) {
                // If stopped, re-queue FlowFile instead of sending it
                session.transfer(flowFile);
                continue;
            }
            final String topic = context.getProperty(TOPIC).evaluateAttributeExpressions(flowFile).getValue();
            final String messageKeyField = context.getProperty(MESSAGE_KEY_FIELD).evaluateAttributeExpressions(flowFile).getValue();
            final Map<String, String> attributes = flowFile.getAttributes();
            try {
                session.read(flowFile, new InputStreamCallback() {

                    @Override
                    public void process(final InputStream rawIn) throws IOException {
                        try (final InputStream in = new BufferedInputStream(rawIn)) {
                            final RecordReader reader = readerFactory.createRecordReader(attributes, in, getLogger());
                            final RecordSet recordSet = reader.createRecordSet();
                            final RecordSchema schema = writerFactory.getSchema(attributes, recordSet.getSchema());
                            lease.publish(flowFile, recordSet, writerFactory, schema, messageKeyField, topic);
                        } catch (final SchemaNotFoundException | MalformedRecordException e) {
                            throw new ProcessException(e);
                        }
                    }
                });
            } catch (final Exception e) {
                // The FlowFile will be obtained and the error logged below, when calling publishResult.getFailedFlowFiles()
                lease.getTracker().fail(flowFile, e);
                continue;
            }
        }
        // Complete the send
        final PublishResult publishResult = lease.complete();
        // Transfer any successful FlowFiles.
        final long transmissionMillis = TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - startTime);
        for (FlowFile success : publishResult.getSuccessfulFlowFiles()) {
            final String topic = context.getProperty(TOPIC).evaluateAttributeExpressions(success).getValue();
            final int msgCount = publishResult.getSuccessfulMessageCount(success);
            success = session.putAttribute(success, MSG_COUNT, String.valueOf(msgCount));
            session.adjustCounter("Messages Sent", msgCount, true);
            final String transitUri = KafkaProcessorUtils.buildTransitURI(securityProtocol, bootstrapServers, topic);
            session.getProvenanceReporter().send(success, transitUri, "Sent " + msgCount + " messages", transmissionMillis);
            session.transfer(success, REL_SUCCESS);
        }
        // Transfer any failures.
        for (final FlowFile failure : publishResult.getFailedFlowFiles()) {
            final int successCount = publishResult.getSuccessfulMessageCount(failure);
            if (successCount > 0) {
                getLogger().error("Failed to send some messages for {} to Kafka, but {} messages were acknowledged by Kafka. Routing to failure due to {}", new Object[] { failure, successCount, publishResult.getReasonForFailure(failure) });
            } else {
                getLogger().error("Failed to send all message for {} to Kafka; routing to failure due to {}", new Object[] { failure, publishResult.getReasonForFailure(failure) });
            }
            session.transfer(failure, REL_FAILURE);
        }
    }
}