DOI License How to Java 11 Java 14 Java 14

SPARQL Anything

SPARQL Anything is a system for Semantic Web re-engineering that allows users to ... query anything with SPARQL.

Main features:


SPARQL Anything uses a single generic abstraction for all data source formats called Facade-X.


Facade-X is a simplistic meta-model used by SPARQL Anything transformers to generate RDF data from diverse data sources. Intuitively, Facade-X uses a subset of RDF as a general approach to represent the source content as-it-is but in RDF. The model combines two types of elements: containers and literals. Facade-X always has a single root container. Container members are a combination of key-value pairs, where keys are either RDF properties or container membership properties. Instead, values can be either RDF literals or other containers. This is a generic example of a Facade-X data object (more examples below):

@prefix fx: <> .
@prefix xyz: <> .
@prefix rdf: <> .
[] a fx:root ; rdf:_1 [
    xyz:someKey "some value" ;
    rdf:_1 "another value with unspecified key" ;
    rdf:_2 [
        rdf:type xyz:MyType ;
        rdf:_1 "another value"
] .

Querying anything

SPARQL Anything extends the Apache Jena ARQ processors by overloading the SERVICE operator, as in the following example:

Suppose having this JSON file as input (also available at

    "summary":"Follows the personal and professional lives of six twenty to thirty-something-year-old friends living in Manhattan.",
      "Jennifer Aniston",
      "Courteney Cox",
      "Lisa Kudrow",
      "Matt LeBlanc",
      "Matthew Perry",
      "David Schwimmer"
    "name":"Cougar Town",
    "summary":"Jules is a recently divorced mother who has to face the unkind realities of dating in a world obsessed with beauty and youth. As she becomes older, she starts discovering herself.",
      "Courteney Cox",
      "David Arquette",
      "Bill Lawrence",
      "Linda Videtti Figueiredo",
      "Blake McCormick"

With SPARQL Anything you can select the TV series starring "Courteney Cox" with the SPARQL query

PREFIX xyz: <>
PREFIX rdf: <>
PREFIX fx: <>

SELECT ?seriesName

    SERVICE <x-sparql-anything:> {
        ?tvSeries xyz:name ?seriesName .
        ?tvSeries xyz:stars ?star .
        ?star fx:anySlot "Courteney Cox" .


and get this result without caring of transforming JSON to RDF.

"Cougar Town"

Using the Command Line Interface

SPARQL Anything requires Java >= 11 to be installed in your operating system. Download the latest version of the SPARQL Anything command line from the releases page. The command line is a file named sparql-anything-<version>.jar. Prepare a file with the query above and name it, for example query.sparql. The query can be executed as follows:

java -jar sparql-anything-<version>.jar -q query.sparql

See the usage section for details on the command line interface.

Using the server

SPARQL Anything is also released as a server, embedded into an instance of the Apache Jena Fuseki server. The server requires Java >= 11 to be installed in your operating system. Download the latest version of the SPARQL Anything server from the releases page. The command line is a file named sparql-anything-server-<version>.jar.

Run the server as follows:

$ java -jar sparql-anything-server-<version>.jar 
[main] INFO com.github.sparqlanything.fuseki.Endpoint - sparql.anything endpoint
[main] INFO com.github.sparqlanything.fuseki.Endpoint - Starting sparql.anything endpoint..
[main] INFO com.github.sparqlanything.fuseki.Endpoint - The server will be listening on http://localhost:3000/sparql.anything
[main] INFO com.github.sparqlanything.fuseki.Endpoint - The server will be available on http://localhost:3000/sparql
[main] INFO org.eclipse.jetty.server.Server - jetty-10.0.6; built: 2021-06-29T15:28:56.259Z; git: 37e7731b4b142a882d73974ff3bec78d621bd674; jvm 11.0.10+9
[main] INFO org.eclipse.jetty.server.handler.ContextHandler - Started o.e.j.s.ServletContextHandler@782a4fff{org.apache.jena.fuseki.Servlet,/,null,AVAILABLE}
[main] INFO org.eclipse.jetty.server.AbstractConnector - Started ServerConnector@c7a975a{HTTP/1.1, (http/1.1)}{}
[main] INFO org.eclipse.jetty.server.Server - Started Server@35beb15e{STARTING}[10.0.6,sto=0] @889ms
[main] INFO org.apache.jena.fuseki.Server - Start Fuseki (http=3000)

Access the SPARQL UI at the address http://localhost:3000/sparql, where you can copy the query above and execute it. See the usage section for details on the SPARQL Anything Fuseki server.

Supported Formats

Currently, SPARQL Anything supports the following list of formats but the possibilities are limitless! The data is interpreted as in the following examples (using default settings).

A detailed description of the interpretation can be found in the following pages:

... and, of course, the triples generated from the these formats can be integrated with the content of any RDF Static file


SPARQL Anything behaves as a standard SPARQL query engine. For example, the SPARQL Anything server will act as a virtual endpoint that can be queried exactly as a remote SPARQL endpoint. In addition, SPARQL Anything provides a rich Command Line Interface (CLI). For information for how to run SPARQL Anything, please see the quickstart and usage sections of the documentation.

Passing triplification options via SERVICE IRI

In order to instruct the query processor to delegate the execution to SPARQL Anything, you can use the following IRI-schema within SERVICE clauses.

x-sparql-anything ':' ([option] ('=' [value])? ','?)+

A minimal URI that uses only the resource locator is also possible.

x-sparql-anything ':' URL

In this case SPARQL Anything guesses the data source type from the file extension.

Note: Use the file:// protocol to reference local files

Passing triplification options via Basic Graph Pattern

Alternatively, options can be provided as basic graph pattern inside the SERVICE clause as follows

PREFIX xyz: <>
PREFIX rdf: <>
PREFIX fx: <>

SELECT ?seriesName

    SERVICE <x-sparql-anything:> {
        fx:properties fx:location "" .
        ?tvSeries xyz:name ?seriesName .
        ?tvSeries xyz:stars ?star .
        ?star fx:anySlot "Courteney Cox" .


Note that

  1. The SERVICE IRI scheme must be x-sparql-anything:.
  2. Each triplification option to pass to the engine corresponds to a triple of the Basic Graph Pattern inside the SERVICE clause.
  3. Such triples must have fx:properties as subject, fx:[OPTION-NAME] as predicate, and a literal or a variable as object.

You can also mix the two modalities as follows.

PREFIX xyz: <>
PREFIX rdf: <>
PREFIX fx: <>

SELECT ?seriesName

    SERVICE <x-sparql-anything:blank-nodes=false> {
        fx:properties fx:location "" .
        ?tvSeries xyz:name ?seriesName .
        ?tvSeries xyz:stars ?star .
        ?star fx:anySlot "Courteney Cox" .


General purpose options

Option name Description Valid Values Default Value
location* The URL of the data source. Any valid URL or (absolute or relative) path of the file system. *
content* The content to be transformed. Any valid literal. *
command* An external command line to be executed. The output is handled according to the option 'media-type' Any valid literal. *
from-archive The filename of the resource to be triplified within an archive. Any filename. No value
root The IRI of generated root resource. Any valid IRI. location + '#' (in case of location argument is set) or '' + md5Hex(content) + '#' (in case of content argument set)
media-type The media-type of the data source. Any valid Media-Type. Supported media types are specified in the pages dedicated to the supported formats No value (the media-type will be guessed from the the file extension)
namespace The namespace prefix for the properties that will be generated. Any valid namespace prefix.
blank-nodes It tells SPARQL Anything to generate blank nodes or not. true/false true
trim-strings Trim all string literals. true/false false
null-string Do not produce triples where the specified string would be in the object position of the triple. Any string No value
http.* A set of options for customising HTTP request method, headers, querystring, and others. More details on the HTTP request configuration No value
triplifier It forces SPARQL Anything to use a specific triplifier for transforming the data source A canonical name of a Java class No value
charset The charset of the data source. Any charset. UTF-8
metadata It tells SPARQL Anything to extract metadata from the data source and to store it in the named graph with URI <> More details true/false false
ondisk It tells SPARQL Anything to use an on disk graph (instead of the default in memory graph). The string should be a path to a directory where the on disk graph will be stored. Using an on disk graph is almost always slower (than using the default in memory graph) but with it you can triplify large files without running out of memory. A path to a directory No value
ondisk.reuse When using an on disk graph, it tells SPARQL Anything to reuse the previous on disk graph. true/false true
strategy The execution strategy. 0 = in memory, all triples; 1 = in memory, only triples matching any of the triple patterns in the where clause 0,1 1
slice The resources is sliced and the SPARQL query executed on each one of the parts. Supported by: CSV (row by row); JSON (when array slice by item, when json object requires json.path); XML (requires xml.path) true/false false
use-rdfs-member It tells SPARQL Anything to use the (super)property rdfs:member instead of container membership properties (rdf:_1, rdf:_2 ...) true/false false

* It is mandatory to provide either location, content, or command.

More details on configuration

Query templates and variable bindings (CLI only)

The SPARQL Anything CLI supports parametrised queries. SPARQL Anything uses the BASIL convention for variable names in queries.

The syntax is based on the underscore character: '_', and can be easily learned by examples:

  • ?_name The variable specifies the API mandatory parameter name. The value is incorporated in the query as plain literal.
  • ?__name The parameter name is optional.
  • ?_name_iri The variable is substituted with the parameter value as a IRI.
  • ?_name_en The parameter value is considered as literal with the language 'en' (e.g., en,it,es, etc.).
  • ?_name_integer The parameter value is considered as literal and the XSD datatype 'integer' is added during substitution.
  • ?_name_prefix_datatype The parameter value is considered as literal and the datatype 'prefix:datatype' is added during substitution. The prefix must be specified according to the SPARQL syntax.

Variable bindings can be passed in two ways via the CLI argument -v|--values:

  • Inline arguments, e.g.: -v paramName=value1 -v paramName=value2 -v paramName2=other
  • Passing an SPARQL Result Set file, e.g.: -v selectResult.xml

In the first case, the engine computes the cardinal product of all the variables bindings included and execute the query for each one of the resulting set of bindings.

In the second case, the query is executed for each set of bindings in the result set.

The following is an example of how parameter can be used in a query:

PREFIX xyz: <>
PREFIX rdf: <>
PREFIX fx: <>

SELECT ?seriesName
    SERVICE <x-sparql-anything:> {
        ?tvSeries xyz:name ?seriesName .
        ?tvSeries xyz:stars ?star .
        ?star fx:anySlot ?_starName .


The value of ?_starName can be passed via the CLI as follows:

java -jar sparql-anything-<version>.jar -q query.sparql -v starName="Courteney Cox"

Functions and magic properties

SPARQL Anything provides a number of magical functions and properties to facilitate the users in querying the sources and constructing knowledge graphs.

NOTE: SPARQL Anything is built on Apache Jena, see a list of supported functions on the Apache Jena documentation.

Name Function/Magic Property Input Output Description
fx:anySlot Magic Property - - This property matches the RDF container membership properties (e.g. rdf:_1, rdf:_2 ...).
fx:cardinal(?a) Function Container membership property Integer fx:cardinal(?a) returns the corresponding cardinal integer from ?a (rdf:_24 -> 24)
fx:before(?a, ?b) Function Container membership properties Boolean fx:before(?a, ?b) returns true if ?a and ?b are container membership properties and ?a is lower than ?b, false otherwise
fx:after(?a, ?b) Function Container membership properties Boolean fx:after(?a, ?b) returns true if ?a and ?b are container membership properties and ?a is higher than ?b, false otherwise
fx:previous(?a) Function Container membership property Container membership property fx:previous(?a) returns the container membership property that preceeds ?a (rdf:_2 -> rdf:_1)
fx:next(?b) Function Container membership property Container membership property fx:next(?b) returns the container membership property that succeedes ?b (rdf:_1 -> rdf:_2)
fx:forward(?a, ?b) Function Container membership property, Integer Container membership property fx:forward(?a, ?b) returns the container membership property that follows ?a of ?b steps (rdf:_2, 5 -> rdf:_7)
fx:backward(?a, ?b) Function Container membership property, Integer Container membership property fx:backward(?a, ?b) returns the container membership property that preceeds ?a of ?b steps (rdf:_24, 4 -> rdf:_20)
fx:String.startsWith(?stringA, ?stringB) Function String, String Boolean fx:String.startsWith wraps java.lang.String.startsWith
fx:String.endsWith(?stringA, ?stringB) Function String, String Boolean fx:String.endsWith wraps java.lang.String.endsWith
fx:String.indexOf(?stringA, ?stringB) Function String, String Integer fx:String.indexOf wraps java.lang.String.indexOf
fx:String.substring(?string) Function String, Integer, (Integer?) String fx:String.substring wraps java.lang.String.substring
fx:String.toLowerCase(?string) Function String String fx:String.toLowerCase wraps java.lang.String.toLowerCase
fx:String.toUpperCase Function String String fx:String.toUpperCase wraps java.lang.String.toUpperCase
fx:String.replace(?string, ?characterA, ?characterB) Function String, Character, Character String fx:String.replace wraps java.lang.String.replace
fx:String.trim(?string) Function String String fx:String.trim wraps java.lang.String.trim
fx:String.stripLeading(?string) Function String String fx:String.stripLeading wraps java.lang.String.stripLeading
fx:String.stripTrailing(?string) Function String String fx:String.stripTrailing wraps java.lang.String.stripTrailing
fx:String.removeTags(?string) Function String String fx:String.removeTags removes the XML tags from the input string
fxWordUtils.capitalize(?string) Function String String WordUtils.capitalize wraps org.apache.commons.text.WordUtils.capitalize
fxWordUtils.capitalizeFully(?string) Function String String fx:WordUtils.capitalizeFully wraps org.apache.commons.text.WordUtils.capitalizeFully
fx:WordUtils.initials(?string) Function String String fx:WordUtils.initials wraps org.apache.commons.text.WordUtils.initials
fx:WordUtils.swapCase(?string) Function String String fx:WordUtils.swapCase wraps org.apache.commons.text.WordUtils.swapCase
fx:WordUtils.uncapitalize(?string) Function String String fx:WordUtils.uncapitalize wraps org.apache.commons.text.WordUtils.uncapitalize
fx:DigestUtils.md2Hex(?string) Function String String fx:DigestUtils.md2Hex wraps org.apache.commons.codec.digest.DigestUtils.md2Hex
fx:DigestUtils.md5Hex(?string) Function String String fx:DigestUtils.md5Hex wraps org.apache.commons.codec.digest.DigestUtils.md5Hex
fx:DigestUtils.sha1Hex(?string) Function String String fx:DigestUtils.sha1Hex wraps org.apache.commons.codec.digest.DigestUtils.sha1Hex
fx:DigestUtils.sha256Hex(?string) Function String String fx:DigestUtils.sha256Hex wraps org.apache.commons.codec.digest.DigestUtils.sha256Hex
fx:DigestUtils.sha384Hex(?string) Function String String fx:DigestUtils.sha384Hex wraps org.apache.commons.codec.digest.DigestUtils.sha384Hex
fx:DigestUtils.sha512Hex(?string) Function String String fx:DigestUtils.sha512Hex wraps org.apache.commons.codec.digest.DigestUtils.sha512Hex
fx:URLEncoder.encode(?string) Function String, String String fx:URLEncoder.encode wraps
fx:URLDecoder.decode(?string) Function String, String String fx:URLDecoder.decode wraps
fx:serial(?a ... ?n) Function Any sequence of nodes Integer The function fx:serial (?a ... ?n) generates an incremental number using the arguments as reference counters. For example, calling fx:serial("x") two times will generate 1 and then 2. Instead, calling fx:serial(?x) multiple times will generate sequential numbers for each value of ?x.
fx:entity(?a ... ?n) Function Any sequence of node URI node The function fx:entity (?a ... ?n) accepts a list of arguments and performs concatenation and automatic casting to string. Container membership properties (rdf:_1,rdf:_2,...) are cast to numbers and then to strings ("1","2").
fx:literal(?a, ?b) Function String, (URI or language code) Literal node The function fx:literal( ?a , ?b ) builds a literal from the string representation of ?a, using ?b either as a typed literal (if a IRI is given) or a lang code (if a string of length of two is given).
fx:bnode(?a) Function Any node Blank node The function fx:bnode( ?a) builds a blank node enforcing the node value as local identifier. This is useful when multiple construct templates are populated with bnode generated on different query solutions but we want them to be joined in the output RDF graph. Apparently, the standard function BNODE does generate a new node for each query solution (see issue #273 for an explanatory case).


Command Line Interface (CLI)

An executable JAR can be obtained from the Releases page.

The jar can be executed as follows:

usage: java -jar sparql.anything-<version>  -q query [-f <output
            format>] [-v <filepath | name=value> ... ]  [-l path] [-o
 -q,--query <query>                    The path to the file storing the
                                       query to execute or the query
 -o,--output <file>                    OPTIONAL - The path to the output
                                       file. [Default: STDOUT]
 -e,--explain                          OPTIONAL - Explain query execution
 -l,--load <load>                      OPTIONAL - The path to one RDF file
                                       or a folder including a set of
                                       files to be loaded. When present,
                                       the data is loaded in memory and
                                       the query executed against it.
 -f,--format <string>                  OPTIONAL -  Format of the output
                                       file. Supported values: JSON, XML,
                                       CSV, TEXT, TTL, NT, NQ. [Default:
                                       TEXT or TTL]
 -s,--strategy <strategy>              OPTIONAL - Strategy for query
                                       evaluation. Possible values: '1' -
                                       triple filtering (default), '0' -
                                       triplify all data. The system
                                       fallbacks to '0' when the strategy
                                       is not implemented yet for the
                                       given resource type.
 -p,--output-pattern <outputPattern>   OPTIONAL - Output filename pattern,
                                       e.g. 'myfile-?friendName.json'.
                                       Variables should start with '?' and
                                       refer to bindings from the input
                                       file. This option can only be used
                                       in combination with 'input' and is
                                       ignored otherwise. This option
                                       overrides 'output'.
 -v,--values <values>                  OPTIONAL - Values passed as input
                                       parameter to a query template. When
                                       present, the query is pre-processed
                                       by substituting variable names with
                                       the values provided. The argument
                                       can be used in two ways. (1)
                                       Providing a single SPARQL ResultSet
                                       file. In this case, the query is
                                       executed for each set of bindings
                                       in the input result set. Only 1
                                       file is allowed. (2) Named variable
                                       bindings: the argument value must
                                       follow the syntax:
                                       var_name=var_value. The argument
                                       can be passed multiple times and
                                       the query repeated for each set of
 -i,--input <input>                    [Deprecated] OPTIONAL - The path to
                                       a SPARQL result set file to be used
                                       as input. When present, the query
                                       is pre-processed by substituting
                                       variable names with values from the
                                       bindings provided. The query is
                                       repeated for each set of bindings
                                       in the input result set.

Logging can be configured adding the following option (SLF4J):



An executable JAR of a SPARQL-Anything-powered Fuseki endpoint can be obtained from the Releases page.

The jar can be executed as follows:

usage: java -jar sparql-anything-server-<version>.jar [-p port] [-e
            sparql-endpoint-path] [-g endpoint-gui-path]
 -e,--path <path>   The path where the server will be running on (Default
 -g,--gui <gui>     The path of the SPARQL endpoint GUI (Default /sparql).
 -p,--port <port>   The port where the server will be running on (Default
                    3000 ).

Also a docker image can be used by following the instructions here.


SPARQL Anything is distributed under Apache 2.0 License

How to cite our work

For citing SPARQL Anything in academic papers please use:

Luigi Asprino, Enrico Daga, Aldo Gangemi, and Paul Mulholland. 2022. Knowledge Graph Construction with a façade: a unified method to access heterogeneous data sources on the Web. ACM Trans. Internet Technol. Just Accepted (2022). Preprint

author = {Asprino, Luigi and Daga, Enrico and Gangemi, Aldo and Mulholland, Paul},
title = {Knowledge Graph Construction with a Fa\c{c}ade: A Unified Method to Access Heterogeneous Data Sources on the Web},
year = {2022},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
issn = {1533-5399},
url = {},
doi = {10.1145/3555312},
abstract = {Data integration is the dominant use case for RDF Knowledge Graphs. However, Web resources come in formats with weak semantics (for example CSV and JSON), or formats specific to a given application (for example BibTex, HTML, and Markdown). To solve this problem, Knowledge Graph Construction (KGC) is gaining momentum due to its focus on supporting users in transforming data into RDF. However, using existing KGC frameworks result in complex data processing pipelines, which mix structural and semantic mappings, whose development and maintenance constitute a significant bottleneck for KG engineers. Such frameworks force users to rely on different tools, sometimes based on heterogeneous languages, for inspecting sources, designing mappings, and generating triples, thus making the process unnecessarily complicated. We argue that it is possible and desirable to equip KG engineers with the ability of interacting with Web data formats by relying on their expertise in RDF and the well-established SPARQL query language&nbsp;[2]. In this article, we study a unified method for data access to heterogeneous data sources with Facade-X, a meta-model implemented in a new data integration system called SPARQL Anything. We demonstrate that our approach is theoretically sound, since it allows a single meta-model, based on RDF, to represent data from (a) any file format expressible in BNF syntax, as well as (b) any relational database. We compare our method to state-of-the-art approaches in terms of usability (cognitive complexity of the mappings) and general performance. Finally, we discuss the benefits and challenges of this novel approach by engaging with the reference user community.},
journal = {ACM Trans. Internet Technol.},
keywords = {RDF, SPARQL, Meta-model, Re-engineering}

Conference paper mainly focussing on system requirements:

Daga, Enrico; Asprino, Luigi; Mulholland, Paul and Gangemi, Aldo (2021). Facade-X: An Opinionated Approach to SPARQL Anything. In: Alam, Mehwish; Groth, Paul; de Boer, Victor; Pellegrini, Tassilo and Pandit, Harshvardhan J. eds. Volume 53: Further with Knowledge Graphs, Volume 53. IOS Press, pp. 58–73.


          volume = {53},
           month = {August},
          author = {Enrico Daga and Luigi Asprino and Paul Mulholland and Aldo Gangemi},
       booktitle = {Volume 53: Further with Knowledge Graphs},
          editor = {Mehwish Alam and Paul Groth and Victor de Boer and Tassilo Pellegrini and Harshvardhan J. Pandit},
           title = {Facade-X: An Opinionated Approach to SPARQL Anything},
       publisher = {IOS Press},
            year = {2021},
         journal = {Studies on the Semantic Web},
           pages = {58--73},
        keywords = {SPARQL; meta-model; re-engineering},
             url = {},
        abstract = {The Semantic Web research community understood since its beginning how crucial it is to equip practitioners with methods to transform non-RDF resources into RDF. Proposals focus on either engineering content transformations or accessing non-RDF resources with SPARQL. Existing solutions require users to learn specific mapping languages (e.g. RML), to know how to query and manipulate a variety of source formats (e.g. XPATH, JSON-Path), or to combine multiple languages (e.g. SPARQL Generate). In this paper, we explore an alternative solution and contribute a general-purpose meta-model for converting non-RDF resources into RDF: {\ensuremath{<}}i{\ensuremath{>}}Facade-X{\ensuremath{<}}/i{\ensuremath{>}}. Our approach can be implemented by overriding the SERVICE operator and does not require to extend the SPARQL syntax. We compare our approach with the state of art methods RML and SPARQL Generate and show how our solution has lower learning demands and cognitive complexity, and it is cheaper to implement and maintain, while having comparable extensibility and efficiency.}