use of org.apache.drill.exec.vector.complex.writer.Float8Writer in project drill by axbaretto.
the class DrillParquetGroupConverter method getConverterForType.
@SuppressWarnings("resource")
private PrimitiveConverter getConverterForType(String name, PrimitiveType type) {
switch(type.getPrimitiveTypeName()) {
case INT32:
{
if (type.getOriginalType() == null) {
IntWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).integer() : mapWriter.integer(name);
return new DrillIntConverter(writer);
}
switch(type.getOriginalType()) {
case UINT_8:
case UINT_16:
case UINT_32:
case INT_8:
case INT_16:
case INT_32:
{
IntWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).integer() : mapWriter.integer(name);
return new DrillIntConverter(writer);
}
case DECIMAL:
{
ParquetReaderUtility.checkDecimalTypeEnabled(options);
Decimal9Writer writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).decimal9() : mapWriter.decimal9(name, type.getDecimalMetadata().getScale(), type.getDecimalMetadata().getPrecision());
return new DrillDecimal9Converter(writer, type.getDecimalMetadata().getPrecision(), type.getDecimalMetadata().getScale());
}
case DATE:
{
DateWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).date() : mapWriter.date(name);
switch(containsCorruptedDates) {
case META_SHOWS_CORRUPTION:
return new DrillCorruptedDateConverter(writer);
case META_SHOWS_NO_CORRUPTION:
return new DrillDateConverter(writer);
case META_UNCLEAR_TEST_VALUES:
return new CorruptionDetectingDateConverter(writer);
default:
throw new DrillRuntimeException(String.format("Issue setting up parquet reader for date type, " + "unrecognized date corruption status %s. See DRILL-4203 for more info.", containsCorruptedDates));
}
}
case TIME_MILLIS:
{
TimeWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).time() : mapWriter.time(name);
return new DrillTimeConverter(writer);
}
default:
{
throw new UnsupportedOperationException("Unsupported type: " + type.getOriginalType());
}
}
}
case INT64:
{
if (type.getOriginalType() == null) {
BigIntWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).bigInt() : mapWriter.bigInt(name);
return new DrillBigIntConverter(writer);
}
switch(type.getOriginalType()) {
case UINT_64:
case INT_64:
{
BigIntWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).bigInt() : mapWriter.bigInt(name);
return new DrillBigIntConverter(writer);
}
case DECIMAL:
{
ParquetReaderUtility.checkDecimalTypeEnabled(options);
Decimal18Writer writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).decimal18() : mapWriter.decimal18(name, type.getDecimalMetadata().getScale(), type.getDecimalMetadata().getPrecision());
return new DrillDecimal18Converter(writer, type.getDecimalMetadata().getPrecision(), type.getDecimalMetadata().getScale());
}
case TIMESTAMP_MILLIS:
{
TimeStampWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).timeStamp() : mapWriter.timeStamp(name);
return new DrillTimeStampConverter(writer);
}
default:
{
throw new UnsupportedOperationException("Unsupported type " + type.getOriginalType());
}
}
}
case INT96:
{
// TODO: replace null with TIMESTAMP_NANOS once parquet support such type annotation.
if (type.getOriginalType() == null) {
if (options.getOption(ExecConstants.PARQUET_READER_INT96_AS_TIMESTAMP).bool_val) {
TimeStampWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).timeStamp() : mapWriter.timeStamp(name);
return new DrillFixedBinaryToTimeStampConverter(writer);
} else {
VarBinaryWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).varBinary() : mapWriter.varBinary(name);
return new DrillFixedBinaryToVarbinaryConverter(writer, ParquetColumnMetadata.getTypeLengthInBits(type.getPrimitiveTypeName()) / 8, mutator.getManagedBuffer());
}
}
}
case FLOAT:
{
Float4Writer writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).float4() : mapWriter.float4(name);
return new DrillFloat4Converter(writer);
}
case DOUBLE:
{
Float8Writer writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).float8() : mapWriter.float8(name);
return new DrillFloat8Converter(writer);
}
case BOOLEAN:
{
BitWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).bit() : mapWriter.bit(name);
return new DrillBoolConverter(writer);
}
case BINARY:
{
if (type.getOriginalType() == null) {
VarBinaryWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).varBinary() : mapWriter.varBinary(name);
return new DrillVarBinaryConverter(writer, mutator.getManagedBuffer());
}
switch(type.getOriginalType()) {
case UTF8:
{
VarCharWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).varChar() : mapWriter.varChar(name);
return new DrillVarCharConverter(writer, mutator.getManagedBuffer());
}
case ENUM:
{
VarCharWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).varChar() : mapWriter.varChar(name);
return new DrillVarCharConverter(writer, mutator.getManagedBuffer());
}
// TODO not sure if BINARY/DECIMAL is actually supported
case DECIMAL:
{
ParquetReaderUtility.checkDecimalTypeEnabled(options);
DecimalMetadata metadata = type.getDecimalMetadata();
if (metadata.getPrecision() <= 28) {
Decimal28SparseWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).decimal28Sparse() : mapWriter.decimal28Sparse(name, metadata.getScale(), metadata.getPrecision());
return new DrillBinaryToDecimal28Converter(writer, metadata.getPrecision(), metadata.getScale(), mutator.getManagedBuffer());
} else {
Decimal38SparseWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).decimal38Sparse() : mapWriter.decimal38Sparse(name, metadata.getScale(), metadata.getPrecision());
return new DrillBinaryToDecimal38Converter(writer, metadata.getPrecision(), metadata.getScale(), mutator.getManagedBuffer());
}
}
default:
{
throw new UnsupportedOperationException("Unsupported type " + type.getOriginalType());
}
}
}
case FIXED_LEN_BYTE_ARRAY:
if (type.getOriginalType() == OriginalType.DECIMAL) {
ParquetReaderUtility.checkDecimalTypeEnabled(options);
DecimalMetadata metadata = type.getDecimalMetadata();
if (metadata.getPrecision() <= 28) {
Decimal28SparseWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).decimal28Sparse() : mapWriter.decimal28Sparse(name, metadata.getScale(), metadata.getPrecision());
return new DrillBinaryToDecimal28Converter(writer, metadata.getPrecision(), metadata.getScale(), mutator.getManagedBuffer());
} else {
Decimal38SparseWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).decimal38Sparse() : mapWriter.decimal38Sparse(name, metadata.getScale(), metadata.getPrecision());
return new DrillBinaryToDecimal38Converter(writer, metadata.getPrecision(), metadata.getScale(), mutator.getManagedBuffer());
}
} else if (type.getOriginalType() == OriginalType.INTERVAL) {
IntervalWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).interval() : mapWriter.interval(name);
return new DrillFixedLengthByteArrayToInterval(writer);
} else {
VarBinaryWriter writer = type.getRepetition() == Repetition.REPEATED ? mapWriter.list(name).varBinary() : mapWriter.varBinary(name);
return new DrillFixedBinaryToVarbinaryConverter(writer, type.getTypeLength(), mutator.getManagedBuffer());
}
default:
throw new UnsupportedOperationException("Unsupported type: " + type.getPrimitiveTypeName());
}
}
use of org.apache.drill.exec.vector.complex.writer.Float8Writer in project drill by apache.
the class DrillParquetGroupConverter method getConverterForType.
protected PrimitiveConverter getConverterForType(String name, PrimitiveType type) {
switch(type.getPrimitiveTypeName()) {
case INT32:
{
if (type.getOriginalType() == null) {
return getIntConverter(name, type);
}
switch(type.getOriginalType()) {
case UINT_8:
case UINT_16:
case UINT_32:
case INT_8:
case INT_16:
case INT_32:
{
return getIntConverter(name, type);
}
case DECIMAL:
{
ParquetReaderUtility.checkDecimalTypeEnabled(options);
return getVarDecimalConverter(name, type);
}
case DATE:
{
DateWriter writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).date(), l -> l.list().date()) : getWriter(name, (m, f) -> m.date(f), l -> l.date());
switch(containsCorruptedDates) {
case META_SHOWS_CORRUPTION:
return new DrillCorruptedDateConverter(writer);
case META_SHOWS_NO_CORRUPTION:
return new DrillDateConverter(writer);
case META_UNCLEAR_TEST_VALUES:
return new CorruptionDetectingDateConverter(writer);
default:
throw new DrillRuntimeException(String.format("Issue setting up parquet reader for date type, " + "unrecognized date corruption status %s. See DRILL-4203 for more info.", containsCorruptedDates));
}
}
case TIME_MILLIS:
{
TimeWriter writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).time(), l -> l.list().time()) : getWriter(name, (m, f) -> m.time(f), l -> l.time());
return new DrillTimeConverter(writer);
}
default:
{
throw new UnsupportedOperationException("Unsupported type: " + type.getOriginalType());
}
}
}
case INT64:
{
if (type.getOriginalType() == null) {
return getBigIntConverter(name, type);
}
switch(type.getOriginalType()) {
case UINT_64:
case INT_64:
return getBigIntConverter(name, type);
case TIMESTAMP_MICROS:
{
TimeStampWriter writer = getTimeStampWriter(name, type);
return new DrillTimeStampMicrosConverter(writer);
}
case TIME_MICROS:
{
TimeWriter writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).time(), l -> l.list().time()) : getWriter(name, MapWriter::time, ListWriter::time);
return new DrillTimeMicrosConverter(writer);
}
case DECIMAL:
{
ParquetReaderUtility.checkDecimalTypeEnabled(options);
return getVarDecimalConverter(name, type);
}
case TIMESTAMP_MILLIS:
{
TimeStampWriter writer = getTimeStampWriter(name, type);
return new DrillTimeStampConverter(writer);
}
default:
{
throw new UnsupportedOperationException("Unsupported type " + type.getOriginalType());
}
}
}
case INT96:
{
// TODO: replace null with TIMESTAMP_NANOS once parquet support such type annotation.
if (type.getOriginalType() == null) {
if (options.getOption(ExecConstants.PARQUET_READER_INT96_AS_TIMESTAMP).bool_val) {
TimeStampWriter writer = getTimeStampWriter(name, type);
return new DrillFixedBinaryToTimeStampConverter(writer);
} else {
VarBinaryWriter writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).varBinary(), l -> l.list().varBinary()) : getWriter(name, (m, f) -> m.varBinary(f), listWriter -> listWriter.varBinary());
return new DrillFixedBinaryToVarbinaryConverter(writer, ParquetColumnMetadata.getTypeLengthInBits(type.getPrimitiveTypeName()) / 8, mutator.getManagedBuffer());
}
}
}
case FLOAT:
{
Float4Writer writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).float4(), l -> l.list().float4()) : getWriter(name, (m, f) -> m.float4(f), l -> l.float4());
return new DrillFloat4Converter(writer);
}
case DOUBLE:
{
Float8Writer writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).float8(), l -> l.list().float8()) : getWriter(name, (m, f) -> m.float8(f), l -> l.float8());
return new DrillFloat8Converter(writer);
}
case BOOLEAN:
{
BitWriter writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).bit(), l -> l.list().bit()) : getWriter(name, (m, f) -> m.bit(f), l -> l.bit());
return new DrillBoolConverter(writer);
}
case BINARY:
{
LogicalTypeAnnotation.LogicalTypeAnnotationVisitor<PrimitiveConverter> typeAnnotationVisitor = new LogicalTypeAnnotation.LogicalTypeAnnotationVisitor<PrimitiveConverter>() {
@Override
public Optional<PrimitiveConverter> visit(LogicalTypeAnnotation.DecimalLogicalTypeAnnotation decimalLogicalType) {
ParquetReaderUtility.checkDecimalTypeEnabled(options);
return Optional.of(getVarDecimalConverter(name, type));
}
@Override
public Optional<PrimitiveConverter> visit(LogicalTypeAnnotation.StringLogicalTypeAnnotation stringLogicalType) {
return Optional.of(getVarCharConverter(name, type));
}
@Override
public Optional<PrimitiveConverter> visit(LogicalTypeAnnotation.EnumLogicalTypeAnnotation stringLogicalType) {
return Optional.of(getVarCharConverter(name, type));
}
};
Supplier<PrimitiveConverter> converterSupplier = () -> {
VarBinaryWriter writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).varBinary(), l -> l.list().varBinary()) : getWriter(name, MapWriter::varBinary, ListWriter::varBinary);
return new DrillVarBinaryConverter(writer, mutator.getManagedBuffer());
};
return Optional.ofNullable(type.getLogicalTypeAnnotation()).map(typeAnnotation -> typeAnnotation.accept(typeAnnotationVisitor)).flatMap(Function.identity()).orElseGet(converterSupplier);
}
case FIXED_LEN_BYTE_ARRAY:
LogicalTypeAnnotation.LogicalTypeAnnotationVisitor<PrimitiveConverter> typeAnnotationVisitor = new LogicalTypeAnnotation.LogicalTypeAnnotationVisitor<PrimitiveConverter>() {
@Override
public Optional<PrimitiveConverter> visit(LogicalTypeAnnotation.DecimalLogicalTypeAnnotation decimalLogicalType) {
ParquetReaderUtility.checkDecimalTypeEnabled(options);
return Optional.of(getVarDecimalConverter(name, type));
}
@Override
public Optional<PrimitiveConverter> visit(LogicalTypeAnnotation.IntervalLogicalTypeAnnotation intervalLogicalType) {
IntervalWriter writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).interval(), l -> l.list().interval()) : getWriter(name, MapWriter::interval, ListWriter::interval);
return Optional.of(new DrillFixedLengthByteArrayToInterval(writer));
}
};
Supplier<PrimitiveConverter> converterSupplier = () -> {
VarBinaryWriter writer = type.isRepetition(Repetition.REPEATED) ? getWriter(name, (m, f) -> m.list(f).varBinary(), l -> l.list().varBinary()) : getWriter(name, MapWriter::varBinary, ListWriter::varBinary);
return new DrillFixedBinaryToVarbinaryConverter(writer, type.getTypeLength(), mutator.getManagedBuffer());
};
return Optional.ofNullable(type.getLogicalTypeAnnotation()).map(typeAnnotation -> typeAnnotation.accept(typeAnnotationVisitor)).flatMap(Function.identity()).orElseGet(converterSupplier);
default:
throw new UnsupportedOperationException("Unsupported type: " + type.getPrimitiveTypeName());
}
}
use of org.apache.drill.exec.vector.complex.writer.Float8Writer in project drill by axbaretto.
the class HttpdLogRecord method set.
/**
* This method is referenced and called via reflection. This is added as a parsing target for the parser. It will get
* called when the value of a log field is a Double data type.
*
* @param field name of field
* @param value value of field
*/
@SuppressWarnings("unused")
public void set(final String field, final Double value) {
if (value != null) {
final Float8Writer w = doubles.get(field);
if (w != null) {
LOG.trace("Parsed field: {}, as double: {}", field, value);
w.writeFloat8(value);
} else {
LOG.warn("No 'double' writer found for field: {}", field);
}
}
}
use of org.apache.drill.exec.vector.complex.writer.Float8Writer in project drill by axbaretto.
the class HttpdLogRecord method setWildcard.
/**
* This method is referenced and called via reflection. When the parser processes a field like:
* HTTP.URI:request.firstline.uri.query.* where star is an arbitrary field that the parser found this method will be
* invoked. <br>
*
* @param field name of field
* @param value value of field
*/
@SuppressWarnings("unused")
public void setWildcard(final String field, final Double value) {
if (value != null) {
final MapWriter mapWriter = getWildcardWriter(field);
LOG.trace("Parsed wildcard field: {}, as double: {}", field, value);
final Float8Writer w = mapWriter.float8(cleanExtensions.get(field));
w.writeFloat8(value);
}
}
Aggregations