RDF terms in rdflib

Terms are the kinds of objects that can appear in a quoted/asserted triples. Those that are part of core RDF concepts are: Blank Node, URI Reference and Literal, the latter consisting of a literal value and either a datatype or an RFC 3066 language tag.

All terms in RDFLib are sub-classes of the rdflib.term.Identifier class.

Nodes are a subset of the Terms that the underlying store actually persists. The set of such Terms depends on whether or not the store is formula-aware. Stores that aren’t formula-aware would only persist those terms core to the RDF Model, and those that are formula-aware would be able to persist the N3 extensions as well. However, utility terms that only serve the purpose for matching nodes by term-patterns probably will only be terms and not nodes.


In RDF, a blank node (also called BNode) is a node in an RDF graph representing a resource for which a URI or literal is not given. The resource represented by a blank node is also called an anonymous resource. By RDF standard a blank node can only be used as subject or object in an RDF triple, although in some syntaxes like Notation 3 [1] it is acceptable to use a blank node as a predicate. If a blank node has a node ID (not all blank nodes are labelled in all RDF serializations), it is limited in scope to a serialization of a particular RDF graph, i.e. the node p1 in the subsequent example does not represent the same node as a node named p1 in any other graph –wikipedia
class rdflib.term.BNode[source]

Blank Node: http://www.w3.org/TR/rdf-concepts/#section-blank-nodes

>>> from rdflib import BNode
>>> anode = BNode()
>>> anode
>>> anode.n3()


A URI reference within an RDF graph is a Unicode string that does not contain any control characters ( #x00 - #x1F, #x7F-#x9F) and would produce a valid URI character sequence representing an absolute URI with optional fragment identifier – W3 RDF Concepts
class rdflib.term.URIRef[source]

RDF URI Reference: http://www.w3.org/TR/rdf-concepts/#section-Graph-URIref

>>> from rdflib import URIRef
>>> aref = URIRef()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: __new__() takes at least 2 arguments (1 given)
>>> aref = URIRef('')
>>> aref
>>> aref = URIRef('http://example.com')
>>> aref
>>> aref.n3()


Literals are the attribute values in RDF, for instance, a person’s name, the date of birth, height, etc. Literals can have a data-type (i.e. this is a double) or a language tag (this label is in English).

class rdflib.term.Literal[source]

RDF Literal: http://www.w3.org/TR/rdf-concepts/#section-Graph-Literal

The lexical value of the literal is the unicode object The interpreted, datatyped value is available from .value

Language tags must be valid according to :rfc:5646

For valid XSD datatypes, the lexical form is optionally normalized at construction time. Default behaviour is set by rdflib.NORMALIZE_LITERALS and can be overridden by the normalize parameter to __new__

Equality and hashing of Literals are done based on the lexical form, i.e.:

>>> from rdflib.namespace import XSD
>>> Literal('01')!=Literal('1') # clear - strings differ

but with data-type they get normalized:

>>> Literal('01', datatype=XSD.integer)!=Literal('1', datatype=XSD.integer)

unless disabled:

>>> Literal('01', datatype=XSD.integer, normalize=False)!=Literal('1', datatype=XSD.integer)

Value based comparison is possible:

>>> Literal('01', datatype=XSD.integer).eq(Literal('1', datatype=XSD.float))

The eq method also provides limited support for basic python types:

>>> Literal(1).eq(1) # fine - int compatible with xsd:integer
>>> Literal('a').eq('b') # fine - str compatible with plain-lit
>>> Literal('a', datatype=XSD.string).eq('a') # fine - str compatible with xsd:string
>>> Literal('a').eq(1) # not fine, int incompatible with plain-lit

Greater-than/less-than ordering comparisons are also done in value space, when compatible datatypes are used. Incompatible datatypes are ordered by DT, or by lang-tag. For other nodes the ordering is None < BNode < URIRef < Literal

Any comparison with non-rdflib Node are “NotImplemented” In PY2.X some stable order will be made up by python

In PY3 this is an error.

>>> from rdflib import Literal, XSD
>>> lit2006 = Literal('2006-01-01',datatype=XSD.date)
>>> lit2006.toPython()
datetime.date(2006, 1, 1)
>>> lit2006 < Literal('2007-01-01',datatype=XSD.date)
>>> Literal(datetime.utcnow()).datatype
>>> Literal(1) > Literal(2) # by value
>>> Literal(1) > Literal(2.0) # by value
>>> Literal('1') > Literal(1) # by DT
>>> Literal('1') < Literal('1') # by lexical form
>>> Literal('a', lang='en') > Literal('a', lang='fr') # by lang-tag
>>> Literal(1) > URIRef('foo') # by node-type

The > < operators will eat this NotImplemented and either make up an ordering (py2.x) or throw a TypeError (py3k):

>>> Literal(1).__gt__(2.0)

A literal in an RDF graph contains one or two named components.

All literals have a lexical form being a Unicode string, which SHOULD be in Normal Form C.

Plain literals have a lexical form and optionally a language tag as defined by RFC 3066, normalized to lowercase. An exception will be raised if illegal language-tags are passed to rdflib.term.Literal.__init__().

Typed literals have a lexical form and a datatype URI being an RDF URI reference.


When using the language tag, care must be taken not to confuse language with locale. The language tag relates only to human language text. Presentational issues should be addressed in end-user applications.


The case normalization of language tags is part of the description of the abstract syntax, and consequently the abstract behaviour of RDF applications. It does not constrain an RDF implementation to actually normalize the case. Crucially, the result of comparing two language tags should not be sensitive to the case of the original input. – RDF Concepts and Abstract Syntax

Python support

RDFLib Literals essentially behave like unicode characters with an XML Schema datatype or language attribute.

datatype hierarchy

The class provides a mechanism to both convert Python literals (and their built-ins such as time/date/datetime) into equivalent RDF Literals and (conversely) convert Literals to their Python equivalent. This mapping to and from Python literals is done as follows:

XML Datatype Python type
None None [1]
xsd:time time [2]
xsd:date date
xsd:dateTime datetime
xsd:string None
xsd:normalizedString None
xsd:token None
xsd:language None
xsd:boolean boolean
xsd:decimal Decimal
xsd:integer long
xsd:nonPositiveInteger int
xsd:long long
xsd:nonNegativeInteger int
xsd:negativeInteger int
xsd:int long
xsd:unsignedLong long
xsd:positiveInteger int
xsd:short int
xsd:unsignedInt long
xsd:byte int
xsd:unsignedShort int
xsd:unsignedByte int
xsd:float float
xsd:double float
xsd:base64Binary base64
xsd:anyURI None
rdf:XMLLiteral xml.dom.minidom.Document [3]
rdf:HTML xml.dom.minidom.DocumentFragment
[1]plain literals map directly to value space
[2]Date, time and datetime literals are mapped to Python instances using the isodate package).
[3]this is a bit dirty - by accident the html5lib parser produces DocumentFragments, and the xml parser Documents, letting us use this to decide what datatype when round-tripping.

An appropriate data-type and lexical representation can be found using:


Casts a python datatype to a tuple of the lexical value and a datatype URI (or None)

and the other direction with

rdflib.term._castLexicalToPython(lexical, datatype)[source]

Map a lexical form to the value-space for the given datatype :returns: a python object for the value or None

All this happens automatically when creating Literal objects by passing Python objects to the constructor, and you never have to do this manually.

You can add custom data-types with rdflib.term.bind(), see also examples.custom_datatype