Examples with BackendException org.apache.pig.backend.BackendException used on opensource projects

Example 1 with BackendException

use of org.apache.pig.backend.BackendException in project hive by apache.

the class HCatBaseStorer method putNext.

@Override
public void putNext(Tuple tuple) throws IOException {
    List<Object> outgoing = new ArrayList<Object>(tuple.size());
    int i = 0;
    for (HCatFieldSchema fSchema : computedSchema.getFields()) {
        outgoing.add(getJavaObj(tuple.get(i++), fSchema));
    }
    try {
        writer.write(null, new DefaultHCatRecord(outgoing));
    } catch (InterruptedException e) {
        throw new BackendException("Error while writing tuple: " + tuple, PigHCatUtil.PIG_EXCEPTION_CODE, e);
    }
}

Example 2 with BackendException

use of org.apache.pig.backend.BackendException in project hive by apache.

the class HCatBaseStorer method getJavaObj.

/**
 * Convert from Pig value object to Hive value object
 * This method assumes that {@link #validateSchema(org.apache.pig.impl.logicalLayer.schema.Schema.FieldSchema, org.apache.hive.hcatalog.data.schema.HCatFieldSchema, org.apache.pig.impl.logicalLayer.schema.Schema, org.apache.hive.hcatalog.data.schema.HCatSchema, int)}
 * which checks the types in Pig schema are compatible with target Hive table, has been called.
 */
private Object getJavaObj(Object pigObj, HCatFieldSchema hcatFS) throws HCatException, BackendException {
    try {
        if (pigObj == null)
            return null;
        // The real work-horse. Spend time and energy in this method if there is
        // need to keep HCatStorer lean and go fast.
        Type type = hcatFS.getType();
        switch(type) {
            case BINARY:
                return ((DataByteArray) pigObj).get();
            case STRUCT:
                HCatSchema structSubSchema = hcatFS.getStructSubSchema();
                // Unwrap the tuple.
                List<Object> all = ((Tuple) pigObj).getAll();
                ArrayList<Object> converted = new ArrayList<Object>(all.size());
                for (int i = 0; i < all.size(); i++) {
                    converted.add(getJavaObj(all.get(i), structSubSchema.get(i)));
                }
                return converted;
            case ARRAY:
                // Unwrap the bag.
                DataBag pigBag = (DataBag) pigObj;
                HCatFieldSchema tupFS = hcatFS.getArrayElementSchema().get(0);
                boolean needTuple = tupFS.getType() == Type.STRUCT;
                List<Object> bagContents = new ArrayList<Object>((int) pigBag.size());
                Iterator<Tuple> bagItr = pigBag.iterator();
                while (bagItr.hasNext()) {
                    // If there is only one element in tuple contained in bag, we throw away the tuple.
                    bagContents.add(getJavaObj(needTuple ? bagItr.next() : bagItr.next().get(0), tupFS));
                }
                return bagContents;
            case MAP:
                Map<?, ?> pigMap = (Map<?, ?>) pigObj;
                Map<Object, Object> typeMap = new LinkedHashMap<Object, Object>();
                for (Entry<?, ?> entry : pigMap.entrySet()) {
                    // the value has a schema and not a FieldSchema
                    typeMap.put(// Schema validation enforces that the Key is a String
                    (String) entry.getKey(), getJavaObj(entry.getValue(), hcatFS.getMapValueSchema().get(0)));
                }
                return typeMap;
            case STRING:
            case INT:
            case BIGINT:
            case FLOAT:
            case DOUBLE:
                return pigObj;
            case SMALLINT:
                if ((Integer) pigObj < Short.MIN_VALUE || (Integer) pigObj > Short.MAX_VALUE) {
                    handleOutOfRangeValue(pigObj, hcatFS);
                    return null;
                }
                return ((Integer) pigObj).shortValue();
            case TINYINT:
                if ((Integer) pigObj < Byte.MIN_VALUE || (Integer) pigObj > Byte.MAX_VALUE) {
                    handleOutOfRangeValue(pigObj, hcatFS);
                    return null;
                }
                return ((Integer) pigObj).byteValue();
            case BOOLEAN:
                if (pigObj instanceof String) {
                    if (((String) pigObj).trim().compareTo("0") == 0) {
                        return Boolean.FALSE;
                    }
                    if (((String) pigObj).trim().compareTo("1") == 0) {
                        return Boolean.TRUE;
                    }
                    throw new BackendException("Unexpected type " + type + " for value " + pigObj + " of class " + pigObj.getClass().getName(), PigHCatUtil.PIG_EXCEPTION_CODE);
                }
                return Boolean.parseBoolean(pigObj.toString());
            case DECIMAL:
                BigDecimal bd = (BigDecimal) pigObj;
                DecimalTypeInfo dti = (DecimalTypeInfo) hcatFS.getTypeInfo();
                if (bd.precision() > dti.precision() || bd.scale() > dti.scale()) {
                    handleOutOfRangeValue(pigObj, hcatFS);
                    return null;
                }
                return HiveDecimal.create(bd);
            case CHAR:
                String charVal = (String) pigObj;
                CharTypeInfo cti = (CharTypeInfo) hcatFS.getTypeInfo();
                if (charVal.length() > cti.getLength()) {
                    handleOutOfRangeValue(pigObj, hcatFS);
                    return null;
                }
                return new HiveChar(charVal, cti.getLength());
            case VARCHAR:
                String varcharVal = (String) pigObj;
                VarcharTypeInfo vti = (VarcharTypeInfo) hcatFS.getTypeInfo();
                if (varcharVal.length() > vti.getLength()) {
                    handleOutOfRangeValue(pigObj, hcatFS);
                    return null;
                }
                return new HiveVarchar(varcharVal, vti.getLength());
            case TIMESTAMP:
                DateTime dt = (DateTime) pigObj;
                // toEpochMilli() returns UTC time regardless of TZ
                return Timestamp.ofEpochMilli(dt.getMillis());
            case DATE:
                /**
                 * We ignore any TZ setting on Pig value since java.sql.Date doesn't have it (in any
                 * meaningful way).  So the assumption is that if Pig value has 0 time component (midnight)
                 * we assume it reasonably 'fits' into a Hive DATE.  If time part is not 0, it's considered
                 * out of range for target type.
                 */
                DateTime dateTime = ((DateTime) pigObj);
                if (dateTime.getMillisOfDay() != 0) {
                    handleOutOfRangeValue(pigObj, hcatFS, "Time component must be 0 (midnight) in local timezone; Local TZ val='" + pigObj + "'");
                    return null;
                }
                /*java.sql.Date is a poorly defined API.  Some (all?) SerDes call toString() on it
        [e.g. LazySimpleSerDe, uses LazyUtils.writePrimitiveUTF8()],  which automatically adjusts
          for local timezone.  Date.valueOf() also uses local timezone (as does Date(int,int,int).
          Also see PigHCatUtil#extractPigObject() for corresponding read op.  This way a DATETIME from Pig,
          when stored into Hive and read back comes back with the same value.*/
                return Date.of(dateTime.getYear(), dateTime.getMonthOfYear(), dateTime.getDayOfMonth());
            default:
                throw new BackendException("Unexpected HCat type " + type + " for value " + pigObj + " of class " + pigObj.getClass().getName(), PigHCatUtil.PIG_EXCEPTION_CODE);
        }
    } catch (BackendException e) {
        // provide the path to the field in the error message
        throw new BackendException((hcatFS.getName() == null ? " " : hcatFS.getName() + ".") + e.getMessage(), e);
    }
}

Example 3 with BackendException

use of org.apache.pig.backend.BackendException in project zeppelin by apache.

the class BasePigInterpreter method cancel.

@Override
public void cancel(InterpreterContext context) {
    LOGGER.info("Cancel paragraph:" + context.getParagraphId());
    PigScriptListener listener = listenerMap.get(context.getParagraphId());
    if (listener != null) {
        Set<String> jobIds = listener.getJobIds();
        if (jobIds.isEmpty()) {
            LOGGER.info("No job is started, so can not cancel paragraph:" + context.getParagraphId());
        }
        for (String jobId : jobIds) {
            LOGGER.info("Kill jobId:" + jobId);
            HExecutionEngine engine = (HExecutionEngine) getPigServer().getPigContext().getExecutionEngine();
            try {
                Field launcherField = HExecutionEngine.class.getDeclaredField("launcher");
                launcherField.setAccessible(true);
                Launcher launcher = (Launcher) launcherField.get(engine);
                // It doesn't work for Tez Engine due to PIG-5035
                launcher.killJob(jobId, new Configuration());
            } catch (NoSuchFieldException | BackendException | IllegalAccessException e) {
                LOGGER.error("Fail to cancel paragraph:" + context.getParagraphId(), e);
            }
        }
    } else {
        LOGGER.warn("No PigScriptListener found, can not cancel paragraph:" + context.getParagraphId());
    }
}