org.apache.drill.exec.util

Class DecimalUtility

java.lang.Object

org.apache.drill.exec.util.DecimalUtility

public class DecimalUtility
extends Object

Field Summary

Fields

Modifier and Type	Field and Description
static int	DIGITS_BASE
static int	INTEGER_SIZE
static int	MAX_DIGITS
static int	MAX_DIGITS_BIGINT
static int	MAX_DIGITS_INT

Constructor Summary

Constructors

Constructor and Description
DecimalUtility()

Method Summary

Modifier and Type	Method and Description
static void	checkValueOverflow(BigDecimal value, int desiredPrecision, int desiredScale) Checks that the specified value may be fit into the value with specified desiredPrecision precision and desiredScale scale.
static int	compareVarLenBytes(DrillBuf left, int leftStart, int leftEnd, int leftScale, byte[] right, int rightScale, boolean absCompare) Compares two VarDecimal values, still stored in Drill buffer and byte array
static int	compareVarLenBytes(DrillBuf left, int leftStart, int leftEnd, int leftScale, DrillBuf right, int rightStart, int rightEnd, int rightScale, boolean absCompare) Compares two VarDecimal values, still stored in their respective Drill buffers
static BigDecimal	getBigDecimalFromDense(DrillBuf data, int startIndex, int nDecimalDigits, int scale, int maxPrecision, int width) Returns a BigDecimal object from the dense decimal representation.
static BigDecimal	getBigDecimalFromDrillBuf(io.netty.buffer.ByteBuf data, int startIndex, int nDecimalDigits, int scale, boolean truncateScale) Create a BigDecimal object using the data in the DrillBuf.
static BigDecimal	getBigDecimalFromDrillBuf(DrillBuf bytebuf, int start, int length, int scale) Create a BigDecimal object using the data in the DrillBuf.
static BigDecimal	getBigDecimalFromPrimitiveTypes(int input, int scale) Returns BigDecimal value created from specified integer value with specified scale.
static BigDecimal	getBigDecimalFromPrimitiveTypes(long input, int scale) Returns BigDecimal value created from specified long value with specified scale.
static BigDecimal	getBigDecimalFromSparse(DrillBuf data, int startIndex, int nDecimalDigits, int scale) Create a BigDecimal object using the data in the DrillBuf.
static long	getDecimal18FromBigDecimal(BigDecimal input, int scale) Returns unsigned long value taken from specified BigDecimal input with specified scale
static int	getDecimal9FromBigDecimal(BigDecimal input, int scale) Returns unsigned int value taken from specified BigDecimal input with specified scale.
static int	getDefaultPrecision(TypeProtos.MinorType minorType, int defaultPrecision) Returns default precision for specified TypeProtos.MinorType or returns specified defaultPrecision if TypeProtos.MinorType isn't TypeProtos.MinorType.INT or TypeProtos.MinorType.BIGINT.
static int	getMaxBytesSizeForPrecision(int precision) Returns max length of byte array, required to store value with specified precision.
static void	getSparseFromBigDecimal(BigDecimal input, io.netty.buffer.ByteBuf data, int startIndex, int scale, int nDecimalDigits) Function converts the BigDecimal and stores it in out internal sparse representation
static boolean	isObsoleteDecimalType(TypeProtos.MinorType minorType) Checks that specified decimal minorType is obsolete.
static int	roundUp(int ndigits) Given the number of actual digits this function returns the number of indexes it will occupy in the array of integers which are stored in base 1 billion
static BigDecimal	sqrt(BigDecimal in, int scale) Calculates and returns square root for specified BigDecimal with specified number of digits alter decimal point.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

MAX_DIGITS

public static final int MAX_DIGITS

See Also:

Constant Field Values

MAX_DIGITS_INT

public static final int MAX_DIGITS_INT

See Also:

Constant Field Values

MAX_DIGITS_BIGINT

public static final int MAX_DIGITS_BIGINT

See Also:

Constant Field Values

DIGITS_BASE

public static final int DIGITS_BASE

See Also:

Constant Field Values

INTEGER_SIZE

public static final int INTEGER_SIZE

See Also:

Constant Field Values

Constructor Detail

DecimalUtility

public DecimalUtility()

Method Detail

roundUp

public static int roundUp(int ndigits)

Given the number of actual digits this function returns the number of indexes it will occupy in the array of integers which are stored in base 1 billion

getBigDecimalFromSparse

public static BigDecimal getBigDecimalFromSparse(DrillBuf data,
                                                 int startIndex,
                                                 int nDecimalDigits,
                                                 int scale)

Create a BigDecimal object using the data in the DrillBuf. This function assumes that data is provided in a sparse format.

getBigDecimalFromDrillBuf

public static BigDecimal getBigDecimalFromDrillBuf(DrillBuf bytebuf,
                                                   int start,
                                                   int length,
                                                   int scale)

Create a BigDecimal object using the data in the DrillBuf. This function assumes that data is provided in format supported by BigInteger.

getBigDecimalFromDrillBuf

public static BigDecimal getBigDecimalFromDrillBuf(io.netty.buffer.ByteBuf data,
                                                   int startIndex,
                                                   int nDecimalDigits,
                                                   int scale,
                                                   boolean truncateScale)

Create a BigDecimal object using the data in the DrillBuf. This function assumes that data is provided in a non-dense format It works on both sparse and intermediate representations.

getBigDecimalFromDense

public static BigDecimal getBigDecimalFromDense(DrillBuf data,
                                                int startIndex,
                                                int nDecimalDigits,
                                                int scale,
                                                int maxPrecision,
                                                int width)

Returns a BigDecimal object from the dense decimal representation. First step is to convert the dense representation into an intermediate representation and then invoke getBigDecimalFromDrillBuf() to get the BigDecimal object

getSparseFromBigDecimal

public static void getSparseFromBigDecimal(BigDecimal input,
                                           io.netty.buffer.ByteBuf data,
                                           int startIndex,
                                           int scale,
                                           int nDecimalDigits)

Function converts the BigDecimal and stores it in out internal sparse representation

getDecimal18FromBigDecimal

public static long getDecimal18FromBigDecimal(BigDecimal input,
                                              int scale)

Returns unsigned long value taken from specified BigDecimal input with specified scale

Parameters:

input - BigDecimal with desired value

scale - scale of the value

Returns:

long value taken from specified BigDecimal

getDecimal9FromBigDecimal

public static int getDecimal9FromBigDecimal(BigDecimal input,
                                            int scale)

Returns unsigned int value taken from specified BigDecimal input with specified scale.

Parameters:

input - BigDecimal with desired value

scale - scale of the value

Returns:

int value taken from specified BigDecimal

getBigDecimalFromPrimitiveTypes

public static BigDecimal getBigDecimalFromPrimitiveTypes(int input,
                                                         int scale)

Returns BigDecimal value created from specified integer value with specified scale.

Parameters:

input - integer value to use for creating of BigDecimal

scale - scale for resulting BigDecimal

Returns:

BigDecimal value

getBigDecimalFromPrimitiveTypes

public static BigDecimal getBigDecimalFromPrimitiveTypes(long input,
                                                         int scale)

Returns BigDecimal value created from specified long value with specified scale.

Parameters:

input - long value to use for creating of BigDecimal

scale - scale for resulting BigDecimal

Returns:

BigDecimal value

compareVarLenBytes

public static int compareVarLenBytes(DrillBuf left,
                                     int leftStart,
                                     int leftEnd,
                                     int leftScale,
                                     DrillBuf right,
                                     int rightStart,
                                     int rightEnd,
                                     int rightScale,
                                     boolean absCompare)

Compares two VarDecimal values, still stored in their respective Drill buffers

Parameters:

left - left value Drill buffer

leftStart - start offset of left value

leftEnd - end offset of left value

leftScale - scale of left value

right - right value Drill buffer

rightStart - start offset of right value

rightEnd - end offset of right value

rightScale - scale of right value

absCompare - comparison of absolute values is done iff this is true

Returns:

1 if left > right, 0 if left = right, -1 if left < right. two values that are numerically equal, but with different scales (e.g., 2.00 and 2), are considered equal.

compareVarLenBytes

public static int compareVarLenBytes(DrillBuf left,
                                     int leftStart,
                                     int leftEnd,
                                     int leftScale,
                                     byte[] right,
                                     int rightScale,
                                     boolean absCompare)

Compares two VarDecimal values, still stored in Drill buffer and byte array

Parameters:

left - left value Drill buffer

leftStart - start offset of left value

leftEnd - end offset of left value

leftScale - scale of left value

right - right value byte array

rightScale - scale of right value

absCompare - comparison of absolute values is done iff this is true

Returns:

1 if left > right, 0 if left = right, -1 if left < right. two values that are numerically equal, but with different scales (e.g., 2.00 and 2), are considered equal.

getMaxBytesSizeForPrecision

public static int getMaxBytesSizeForPrecision(int precision)

Returns max length of byte array, required to store value with specified precision.

Parameters:

precision - the precision of value

Returns:

max length of byte array

sqrt

public static BigDecimal sqrt(BigDecimal in,
                              int scale)

Calculates and returns square root for specified BigDecimal with specified number of digits alter decimal point.

Parameters:

in - BigDecimal which square root should be calculated

scale - number of digits alter decimal point in the result value.

Returns:

square root for specified BigDecimal

isObsoleteDecimalType

public static boolean isObsoleteDecimalType(TypeProtos.MinorType minorType)

Checks that specified decimal minorType is obsolete.

Parameters:

minorType - type to check

Returns:

true if specified decimal minorType is obsolete.

getDefaultPrecision

public static int getDefaultPrecision(TypeProtos.MinorType minorType,
                                      int defaultPrecision)

Returns default precision for specified TypeProtos.MinorType or returns specified defaultPrecision if TypeProtos.MinorType isn't TypeProtos.MinorType.INT or TypeProtos.MinorType.BIGINT.

Parameters:

minorType - type wich precision should be received

defaultPrecision - default value for precision

Returns:

default precision for specified TypeProtos.MinorType

checkValueOverflow

public static void checkValueOverflow(BigDecimal value,
                                      int desiredPrecision,
                                      int desiredScale)

Checks that the specified value may be fit into the value with specified desiredPrecision precision and desiredScale scale. Otherwise, the exception is thrown.

Parameters:

value - BigDecimal value to check

desiredPrecision - precision for the resulting value

desiredScale - scale for the resulting value