2d:null 2e:["$","title",null,{"children":"A SPARQL Tour of MaterialsMine | Michael Deagen | Observable"}] 33:T52f,### Assembling a Query SPARQL is a *declarative* language, meaning that you specify *what* you want at a high level. This is different from an *imperative* language, in which you must specify low-level implementation details that determine *how* the query operations are executed and in what order. The advantage of a declarative language is its minimal syntactic overhead, while imperative languages provide more control over performance. To specify a query, create a BGP by enumerating triples. Using the [Turtle](https://en.wikipedia.org/wiki/Turtle_%28syntax%29) syntax is useful to make the queries less verbose and easier to visualize. Because the specification of a BGP does require some prior knowledge of how data are structured in the knowledge graph, examples like these can be helpful. However, you can always add variables to the BGP when you are unsure, and the query will return matches from the KG that exist. This makes it possible to gradually discover all the information in the KG (the "follow-your-nose" phenomenon). For example, the following query (written using Turtle syntax) will return all unique research article DOIs linked to polymer nanocomposite samples containing `"Silicon dioxide"` as a filler material. To only return unique values for the DOI, `DISTINCT` is added to the `SELECT` clause.34:[] 31:{"id":"d622871f45548c92","slug":"a-sparql-tour-of-materialsmine","title":"A SPARQL Tour of MaterialsMine","thumbnail":"029c3e7dd5053e5b65f8ad39c1eee9bbe76ed1895d311687244ef2b7b74ff6e5","access_level":"public","create_time":"$D2021-05-18T13:33:54.052Z","update_time":"$D2021-08-19T13:38:25.781Z","view_time":null,"first_public_time":"$D2021-05-18T20:39:29.607Z","has_unpublished_edits":null,"fork":false,"stars":2,"starred":false,"owner":{"type":"team","id":"5ca8a3812ea4a326","login":"mdeagen","name":"Michael Deagen","avatar_url":"https://avatars-next.observableusercontent.com/0b8e876e2eac1058faa6c1a2624346f3e72350b8a7e6c9acaec6a9ad920b0476","tier":"starter_2024"},"version":381,"publish_version":381,"role":null,"authored":false,"authors":[{"type":"individual","id":"d3611c0325cb99e9","login":"mdeagen","name":"Michael Deagen","avatar_url":"https://avatars-next.observableusercontent.com/0b8e876e2eac1058faa6c1a2624346f3e72350b8a7e6c9acaec6a9ad920b0476","tier":null}],"members":[{"type":"individual","id":"d3611c0325cb99e9","login":"mdeagen","name":"Michael Deagen","avatar_url":"https://avatars-next.observableusercontent.com/0b8e876e2eac1058faa6c1a2624346f3e72350b8a7e6c9acaec6a9ad920b0476","tier":null,"role":"editor"}],"comments":[],"custom_thumbnail":false,"trashed":false,"lapsed":null,"subscribed":null,"forks":3,"collection":{"id":"07bad1e1751e43d5","slug":"documentation","title":"MaterialsMine Docs","owner":{"id":"5ca8a3812ea4a326","login":"mdeagen","name":"Michael Deagen"}},"type":"notebook","body":{"title":"A SPARQL Tour of MaterialsMine","cells":[{"id":0,"value":"# A SPARQL Tour of MaterialsMine\n##### (see Appendix below for additional links and resources)\n\nData in the [MaterialsMine](https://materialsmine.org/) knowledge graph (KG) are structured as **[RDF](https://www.w3.org/TR/rdf11-primer/) triples** that collectively form a directed network of resources (*nodes*) and the relationships (*edges*) between those resources. The *subject* of a triple can be a URI or a blank node, the *predicate* is a URI, and the *object* is a URI, blank node, or literal.\n\n","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":311,"value":"dot`digraph x {\ngraph [rankdir = \"LR\"];\nnode [style=filled];\n\"subject\" -> \"object\" [label=\"predicate\"];\n}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":315,"value":"By linking triples together, you begin to create a machine-interpretable representation of knowledge in a domain area. Structured vocabularies, known as *ontologies*, enable you to perform inferencing. Re-use of other ontologies is highly encouraged because it improves the odds that an RDF database will be *interoperable* with other RDF databases on the semantic web.\n\nFor example, if you wish to assert that a polymer-matrix nanocomposite comprises two component materials, each of which has a role (matrix/filler), you can capture a representation of that information that enables you to link to other information later (assuming you use the same URIs or can otherwise perform entity resolution).","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":322,"value":"dot`digraph x {\nnode [style=filled];\n\"nm_s:sampleID\" -> \"nm:PolymerNanocomposite\" [label=\"rdf:type\"];\n\"nm_s:sampleID\" -> \"bnode01\" [label=\"sio:hasComponentPart\"];\n\"nm_s:sampleID\" -> \"bnode02\" [label=\"sio:hasComponentPart\"];\n\"bnode01\" -> \"nm_c:SiliconDioxide\" [label=\"rdf:type\"];\n\"bnode01\" -> \"nm:Filler\" [label=\"sio:hasRole\"];\n\"bnode02\" -> \"nm_c:Polyethylene\" [label=\"rdf:type\"];\n\"bnode02\" -> \"nm:Matrix\" [label=\"sio:hasRole\"];\n\n\"bnode01\" [shape=diamond,style=filled,label=\"\",height=.2,width=.2];\n\"bnode02\" [shape=diamond,style=filled,label=\"\",height=.2,width=.2];\n\"nm_c:SiliconDioxide\" [fillcolor=lightblue];\n}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":341,"value":"By making assertions about URIs, that information becomes linked without having to restate each occurrence. However, because of the *open-world assumption*, we cannot assume that the knowledge will be complete (or even correct). This is where careful data modeling and curation practices play an important role!","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":353,"value":"dot`digraph x {\nnode [style=filled];\n\"nm_c:SiliconDioxide\" -> \"''Silicon dioxide''\" [label=\"rdfs:label\"];\n\"nm_c:SiliconDioxide\" -> \"bnode01\" [label=\"sio:hasAttribute\"];\n\"bnode01\" -> \"nm:Density\" [label=\"rdf:type\"];\n\"bnode01\" -> \"''2.65''^^xsd:double\" [label=\"sio:hasValue\"];\n\"bnode01\" -> \"qudt:GM-PER-CentiM2\" [label=\"sio:hasUnit\"];\n\"bnode01\" [shape=diamond,style=filled,label=\"\",height=.2,width=.2];\n\"nm_c:SiliconDioxide\" [fillcolor=lightblue];\n}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":378,"value":"A note on the above graph modeling scheme: the small diamond-shaped node is a *blank node*, or **bnode**, which represents an unnamed resource in the graph.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":307,"value":"### What Is SPARQL?\n[**SPARQL**](https://www.w3.org/TR/2013/REC-sparql11-query-20130321/) is the query language for RDF data, and the interface for running queries on a knowledge graph is known as a **SPARQL endpoint**. The SPARQL endpoint for DBpedia is located at [this link](https://dbpedia.org/sparql), and the SPARQL endpoint for MaterialsMine is [here](https://materialsmine.org/wi/sparql.html).\n\nA SPARQL query defines a **basic graph pattern (BGP)** that is matched against the KG, and variable **bindings** that match the BGP are returned as tabular data. Variables in the query have the prefix `?` followed by a string of characters, and each binding is a pair (variable, RDF term) that matches the BGP in the result set.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":44,"value":"### SELECT Queries\nThe most basic **SELECT** query comprises a single triple of three variables, which will return any triple in the KG. Here, the variables are named *?subject*, *?predicate*, *?object*, but could have been given other names.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":121,"value":"

\n SELECT * WHERE {
\n ?subject ?predicate ?object .
\n }
\n LIMIT 10\n

","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":129,"value":"The asterisk (\\*) in the `SELECT` clause indicates that bindings to all variables in the query will be returned. Alternatively, if you were only interested in the bindings to the variables `?subject` and `?object`, you could replace the above `SELECT` clause with `SELECT ?subject ?object WHERE`.\n\nAlso, this example includes a **LIMIT** clause to cease the execution of the query after the first ten bindings are returned. This helps avoid needlessly long query runtimes.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":106,"value":"### Prefixes\n\nThe top of a SPARQL query generally contains a list of **prefixes** that allow us to refer to namespaces without having to type out the full URI each time.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":104,"value":"

\n PREFIX mm: <http://materialsmine.org/ns/>\n

","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":113,"value":"With the prefix above, <http://materialsmine.org/ns/PolymerNanocomposite> can be written as *mm:PolymerNanocomposite*. This is known as a compact URI, or CURIE. The example below shows a simple query that will return ten URIs of polymer nanocomposite samples from the MaterialsMine KG.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":134,"value":"

\n PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
\n PREFIX mm: <http://materialsmine.org/ns/>
\n SELECT * WHERE {
\n ?sample rdf:type nm:PolymerNanocomposite .
\n }
\n LIMIT 10\n

","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":154,"value":"### Useful Namespaces to Know\n\nThe MaterialsMine KG borrows vocabulary from other ontologies on the semantic web, which improves the interoperability of the system. These ontologies include:\n- Semanticscience Integrated Ontology ([SIO](https://bioportal.bioontology.org/ontologies/SIO)) for scientific metadata\n- Provenance Ontology ([PROV-O](https://www.w3.org/TR/prov-o/)) for data provenance\n- Dublin Core ([DC](https://www.dublincore.org/specifications/dublin-core/)) Ontology for bibliographic metadata\n- and others...\n\nIf you come across a namespace you do not recognize, [prefix.cc](http://prefix.cc/) can help.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":162,"value":"$33","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":171,"value":"

\n PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
\n PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
\n PREFIX mm: <http://materialsmine.org/ns/>
\n PREFIX dct: <http://purl.org/dc/terms/>
\n PREFIX sio: <http://semanticscience.org/resource/>
\n SELECT DISTINCT ?doi WHERE {
\n ?doi rdf:type dct:BibliographicResource ;
\n sio:hasPart ?sample .
\n ?sample rdf:type nm:PolymerNanocomposite ;
\n sio:hasComponentPart [ a ?Material ; sio:hasRole ?role ] .
\n ?Material rdfs:label \"Silicon dioxide\" .
\n ?role rdf:type nm:Filler .
\n }\n

","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":200,"value":"The same query can be visualized as the graph layout below.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":204,"value":"dot`digraph x {\nnode [style=filled];\n\"?doi\" -> \"dct:BibliographicResource\" [label=\"rdf:type\"];\n\"?doi\" -> \"?sample\" [label=\"sio:hasPart\"];\n\"?sample\" -> \"nm:PolymerNanocomposite\" [label=\"rdf:type\"];\n\"?sample\" -> \"bnode01\" [label=\"sio:hasComponentPart\"];\n\"bnode01\" [shape=diamond,style=filled,label=\"\",height=.2,width=.2] ;\n\"bnode01\" -> \"?Material\" [label=\"rdf:type\"];\n\"bnode01\" -> \"?role\" [label=\"sio:hasRole\"];\n\"?Material\" -> \"''Silicon dioxide''\" [label=\"rdfs:label\"];\n\"?role\" -> \"nm:Filler\" [label=\"rdf:type\"];\n\"?doi\" [shape=rectangle];\n\"?sample\" [shape=rectangle];\n\"?Material\" [shape=rectangle];\n\"?role\" [shape=rectangle];\n}`","pinned":false,"hidden":false,"mode":"ojs","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":226,"value":"Intermediate variables can be explicitly enumerated (such as `?Material` and `?role`), or implied using square brackets `[ ]` (this node is shown as a diamond above).","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":28,"value":"### Aggregation in SPARQL\n\nInstead of returning all DOIs in the example above, you may just be interested in the total **count** of DOIs. In this case, you may consider the aggregate operation `COUNT`, which will return the total count of bindings that match the BGP.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":235,"value":"

\n ...
\n SELECT (COUNT(DISTINCT(?doi) AS ?CountofDOIs) WHERE {
\n ...
\n }\n

","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":253,"value":"If instead, you wanted to see a breakdown of DOI count for different material types, you would want to return `?Material` as a bound variable. Because such an operation requires the results to be grouped based on a variable, we use the `GROUP BY` clause after the `SELECT` clause to define how to sub-divide the results. Additionally, you can sort the results using an `ORDER BY` or `ORDER BY DESC()` clause.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":259,"value":"

\n ...
\n SELECT ?Material (COUNT(DISTINCT(?doi) AS ?CountofDOIs) WHERE {
\n ...
\n }
\n GROUP BY ?Material
\n ORDER BY DESC (?CountofDOIs)
\n

","pinned":false,"hidden":false,"mode":"html","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":271,"value":"The above query would return two columns, labeled `Material` and `CountofDOIs`, where the rows are sorted in descending order by the count of DOIs linked to that particular material.","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":275,"value":"### Other SPARQL Functions\n\nThis is only a brief introduction to the SPARQL language. Take a look at the [SPARQL documentation](https://www.w3.org/TR/2013/REC-sparql11-query-20130321/) for more information about its capabilities. In particular, check out:\n\n- `FILTER` clauses\n- `VALUES` clauses\n- operators such as `>`, `<`, `&&`, `||`, `=`, `!=`, `>=`, etc.\n- functions on RDF terms such as `STR`, `isLiteral`, and `isNumeric`\n- string functions such as `REGEX`, `LCASE`, and `STRLEN`\n- other aggregate operators such as `GROUP_CONCAT`, `AVG`, and `MAX`","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":292,"value":"### Types of Information in MaterialsMine\n\nWe curate data and metadata about polymer-matrix nanocomposite materials. Our vocabulary includes materials components, their roles in a nanocomposite (Matrix, Filler, or Surface Treatment), processing steps, characterization metadata, bibliographic metadata, and numerous materials properties. Example data visualizations created from SPARQL query results can be browsed in our [Gallery of Charts](https://materialsmine.org/wi/gallery). Click on the

button on a chart page to preview the SPARQL query, which you can copy and try for yourself at the [MaterialsMine SPARQL endpoint](https://materialsmine.org/wi/sparql.html).","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"},{"id":10,"value":"---\n\n## Appendix\n\n- [Glossary of terms & acronyms](https://observablehq.com/@mdeagen/glossary-of-terms-acronyms)\n- [RDF Primer](https://www.w3.org/TR/rdf11-primer/)\n- [MaterialsMine SPARQL endpoint](https://materialsmine.org/wi/sparql.html)\n- [Observable notebook for querying the MaterialsMine SPARQL endpoint](https://observablehq.com/@mdeagen/materialsmine-sparql)\n- [Observable notebook with a SPARQL query function](https://observablehq.com/@mdeagen/sparql-select-query-function)","pinned":false,"hidden":false,"mode":"md","author":"$undefined","summary":"$undefined","output":"$undefined","database":"$undefined"}],"files":"$W34","stdlib":"1","titling":"auto"}} 32:"$31:body"