com.bigdata.rdf.internal

Enum DTE

java.lang.Object

java.lang.Enum<DTE>

com.bigdata.rdf.internal.DTE

All Implemented Interfaces:

Serializable, Comparable<DTE>

public enum DTE
extends Enum<DTE>

Data Type Enumeration (DTE) is a class which declares the known intrinsic data types, provides for extensibility to new data types, and provides for data types which either can not be inlined or are not being inlined. The intrinsic data types are those having natural orders which can be encoded into an unsigned byte[] key and decoded without loss. Whether or not a given data type is actually inlined is a configuration option for the lexicon.

If a data type is not inlined, then the representation of the value must be materialized. Non-inline values are BlobIVs and are materialized by looking IV in the TERMS index. More recently, we are also permitting indirection from the TERMS index. For example, very large objects may be stored in the file system, in S3, etc. In those cases, you must still resolve the IV against the TERMS index. Now the resulting object may be a BigdataValue. However, the object materialized from the TERMS index may also provide indirection into the file system, S3, etc.

The VTE has 4 distinctions (URI, Literal, BlankNode, and SID) and is coded in the high 2 bits of a byte while the DTE has 16 possible distinctions, one of which is reserved against future use and one of which is reserved against extensibility in the set of intrinsic types.

Note: Unicode values CAN NOT be inlined because (a) Unicode sort keys are not decodable; and (b) the collation rules for Unicode depend on the lexicon configuration, which specifies parameters such as Locale, Strength, etc.

Blanks nodes (their IDs are UUIDs) and data types with larger values (UUIDs) or varying length values (xsd:integer, xsd:decimanl) can be inlined. Whether it makes sense to do so is a question which trades off redundancy in the statement indices for faster materialization of the data type values and a smaller lexicon. UUIDs for purposes other than blank nodes can also be inlined, however they will have a different prefix to indicate that they are Literals rather than blank nodes.

Note: While multidimensional data types (such as points or rectangles) could be inlined (in the sense that their values could be converted to unsigned byte[] keys and the keys could be decoded), they can not be placed into a total order which has meaningful semantics by a single index. For example, geo:point describes a 2-dimensional location and lacks any meaningful locality when inlined into the index. The only reason to inline such data types is to avoid indirection through the lexicon to materialize their values. Efficiently resolving points in a region requires the use of a spatial index.

Author:

Bryan Thompson

See Also:

http://www.w3.org/TR/xmlschema-2/

Enum Constant Summary

Enum Constants

Enum Constant and Description
Extension This provides an extension of the intrinsic data types.
UUID The "inline" value is a UUID.
XSDBoolean The "inline" value is a boolean (xsd:boolean).
XSDByte The "inline" value is a signed byte (xsd:byte).
XSDDecimal The "inline" value is an xsd:decimal.
XSDDouble The "inline" value is a double precision floating point number (xsd:double).
XSDFloat The "inline" value is a single precision floating point number (xsd:float).
XSDInt The "inline" value is a signed 4 byte integer (xsd:int).
XSDInteger The "inline" value is an xsd:integer, which is equivalent to BigInteger.
XSDLong The "inline" value is a signed 8 byte integer (xsd:long).
XSDShort The "inline" value is a signed short (xsd:short).
XSDString The "inline" value is a compressed Unicode string.
XSDUnsignedByte The "inline" value is an unsigned byte (xsd:unsignedByte).
XSDUnsignedInt The "inline" value is an unsigned 4 byte integer (xsd:unsignedInt).
XSDUnsignedLong The "inline" value is an unsigned 8 byte integer (xsd:unsignedLong).
XSDUnsignedShort The "inline" value is a unsigned short (xsd:unsignedShort).

Method Summary

Methods

Modifier and Type	Method and Description
Class<?>	getCls() The class of the Java object used to represent instances of the coded data type.
org.openrdf.model.URI	getDatatypeURI() The corresponding datatype URI.
boolean	isBigNumeric() true for xsd:integer and xsd:decimal.
boolean	isFixedNumeric() This is !isBigNumeric() and is true for any of the fixed length numeric data types (xsd:byte, xsd:unsignedByte, xsd:short, xsd:unsignedShort, xsd:int, xsd:unsignedInt, xsd:long, xsd:unsignedLong, xsd:float, xsd:double).
boolean	isFloatingPointNumeric() true for xsd:float, xsd:double, and xsd:decimal.
boolean	isNumeric() true for any of the numeric data types (xsd:byte, xsd:unsignedByte, xsd:short, xsd:unsignedShort, xsd:int, xsd:unsignedInt, xsd:long, xsd:unsignedLong, xsd:float, xsd:double, xsd:integer, and xsd:decimal).
boolean	isSignedNumeric() true for an signed numeric datatype ( xsd:byte, xsd:short, xsd:int, xsd:long, xsd:float, xsd:double, xsd:integer, and xsd:decimal).
boolean	isUnsignedNumeric() true for an unsigned numeric datatype ( xsd:unsignedByte, xsd:unsignedShort, xsd:unsignedInt, xsd:unsignedLong).
int	len() The length of the data type value when represented as a component in an unsigned byte[] key -or- ZERO iff the key component has a variable length for that data type.
byte	v() An byte value whose whose lower 6 bits code the DTE.
static DTE	valueOf(byte b)
static DTE	valueOf(String name) Returns the enum constant of this type with the specified name.
static DTE	valueOf(org.openrdf.model.URI datatype) Return the DTE for the datatype URI.
static DTE[]	values() Returns an array containing the constants of this enum type, in the order they are declared.

Methods inherited from class java.lang.Enum

clone, compareTo, equals, finalize, getDeclaringClass, hashCode, name, ordinal, toString, valueOf

Methods inherited from class java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

Enum Constant Detail

XSDBoolean

public static final DTE XSDBoolean

The "inline" value is a boolean (xsd:boolean). Only the distinct points in the xsd:boolean value space are represented. xsd:boolean has multiple lexical forms which map onto "true" and "false". Those distinctions are not preserved.

XSDByte

public static final DTE XSDByte

The "inline" value is a signed byte (xsd:byte).

XSDShort

public static final DTE XSDShort

The "inline" value is a signed short (xsd:short).

XSDInt

public static final DTE XSDInt

The "inline" value is a signed 4 byte integer (xsd:int).

XSDLong

public static final DTE XSDLong

The "inline" value is a signed 8 byte integer (xsd:long).

XSDUnsignedByte

public static final DTE XSDUnsignedByte

The "inline" value is an unsigned byte (xsd:unsignedByte).

XSDUnsignedShort

public static final DTE XSDUnsignedShort

The "inline" value is a unsigned short (xsd:unsignedShort).

XSDUnsignedInt

public static final DTE XSDUnsignedInt

The "inline" value is an unsigned 4 byte integer (xsd:unsignedInt).

XSDUnsignedLong

public static final DTE XSDUnsignedLong

The "inline" value is an unsigned 8 byte integer (xsd:unsignedLong).

XSDFloat

public static final DTE XSDFloat

The "inline" value is a single precision floating point number (xsd:float).

XSDDouble

public static final DTE XSDDouble

The "inline" value is a double precision floating point number (xsd:double).

XSDInteger

public static final DTE XSDInteger

The "inline" value is an xsd:integer, which is equivalent to BigInteger.

XSDDecimal

public static final DTE XSDDecimal

The "inline" value is an xsd:decimal. This is mostly equivalent to BigDecimal, but unlike that Java class, xsd:decimal DOES NOT preserve the precision of the value. (This fact is convenient for indices since BigDecimal has, among other things, many distinct representations of ZERO with different precision, etc. If we had to represent the precision, we could not use xsd:decimal in an index!)

UUID

public static final DTE UUID

The "inline" value is a UUID.

See Also:

http://lists.xml.org/archives/xml-dev/201003/msg00027.html

XSDString

public static final DTE XSDString

The "inline" value is a compressed Unicode string. This is decodable compressed encoding rather than a Unicode sort key. It is suitable for representing "small" Unicode values directly within the statement indices. "Small" is configurable, but should not be overly large. The tradeoff is the growth in the B+Tree leaf size for a statement index versus the overhead required when indirecting through the LexiconRelation to materialize the RDF Value. Further, there is a practical upper bound on the size of a key in the B+Tree. Therefore, inlining of Unicode values having between 32 and 64 characters is suggested as a recommended practice. Beyond that, inlining can contribute significantly to the growth in the B+Tree leaf size and have a negative impact on join performance for the statement indices.

This DTE may be used in combination with VTE as follows:

VTE.BNODE

Represent an inline blank node identifier.

VTE.LITERAL

Represent a plain literal.

VTE.LITERAL plus the extension bit

Represent a data type literal where the extension IV is the data type of the literal.

VTE.URI plus the extension bit

Represent a URI where the extension IV is the namespace of the URI and the inline Unicode component is the local name of the URI. This depends on the openrdf definition of a URI's namespace and local name (basically, everything after the last '/' in the URI path or after the '#' if there is a URI anchor).

Extension

public static final DTE Extension

This provides an extension of the intrinsic data types. Its code corresponds to 0xf, which is to say all four bits are on. When this code is used, you must interpret the next byte using DTEExtension.

Note: This is NOT the same as the AbstractIV#isExtension() bit. The latter always indicates that an IV follows the flags byte and indicates the actual datatype URI. In contrast, Extension gives you another byte which you can use to handle additional "intrinsic" types.

See Also:

DTEExtension, (Implement support for DTE extension types for URIs)

Method Detail

values

public static DTE[] values()

Returns an array containing the constants of this enum type, in the order they are declared. This method may be used to iterate over the constants as follows:

for (DTE c : DTE.values())
    System.out.println(c);

Returns:

an array containing the constants of this enum type, in the order they are declared

valueOf

public static DTE valueOf(String name)

Returns the enum constant of this type with the specified name. The string must match exactly an identifier used to declare an enum constant in this type. (Extraneous whitespace characters are not permitted.)

Parameters:

name - the name of the enum constant to be returned.

Returns:

the enum constant with the specified name

Throws:

IllegalArgumentException - if this enum type has no constant with the specified name

NullPointerException - if the argument is null

valueOf

public static final DTE valueOf(byte b)

valueOf

public static final DTE valueOf(org.openrdf.model.URI datatype)

Return the DTE for the datatype URI.

Parameters:

datatype - The datatype URI.

Returns:

The DTE for that datatype -or- Extension if the datatype is null (there is no specific datatype for an extension since extensions by their nature can handle any datatype) -or- null if the datatype URI is none of the datatypes for which native support is provided.

v

public final byte v()

An byte value whose whose lower 6 bits code the DTE.

len

public final int len()

The length of the data type value when represented as a component in an unsigned byte[] key -or- ZERO iff the key component has a variable length for that data type.

getCls

public final Class<?> getCls()

The class of the Java object used to represent instances of the coded data type.

getDatatypeURI

public final org.openrdf.model.URI getDatatypeURI()

The corresponding datatype URI.

isNumeric

public boolean isNumeric()

true for any of the numeric data types (xsd:byte, xsd:unsignedByte, xsd:short, xsd:unsignedShort, xsd:int, xsd:unsignedInt, xsd:long, xsd:unsignedLong, xsd:float, xsd:double, xsd:integer, and xsd:decimal).

isSignedNumeric

public boolean isSignedNumeric()

true for an signed numeric datatype ( xsd:byte, xsd:short, xsd:int, xsd:long, xsd:float, xsd:double, xsd:integer, and xsd:decimal).

isUnsignedNumeric

public boolean isUnsignedNumeric()

true for an unsigned numeric datatype ( xsd:unsignedByte, xsd:unsignedShort, xsd:unsignedInt, xsd:unsignedLong).

isFixedNumeric

public boolean isFixedNumeric()

This is !isBigNumeric() and is true for any of the fixed length numeric data types (xsd:byte, xsd:unsignedByte, xsd:short, xsd:unsignedShort, xsd:int, xsd:unsignedInt, xsd:long, xsd:unsignedLong, xsd:float, xsd:double).

isBigNumeric

public boolean isBigNumeric()

true for xsd:integer and xsd:decimal.

isFloatingPointNumeric

public boolean isFloatingPointNumeric()

true for xsd:float, xsd:double, and xsd:decimal.