Data Sources & Adapters

Register a source

registerSource pairs an in-memory value with a name. It does not copy or convert anything; it is a constant-time operation, so you can register large structures freely.

db.registerSource(['users'], usersArray);
db.registerSource(['crm', 'invoices'], invoicesArray); // namespaced name
db.registerSource(['social'], graph);

The name is an array of parts, which lets you namespace sources; a query then refers to the source by the dotted path, for example crm.invoices.

Anything can be registered: an array, an object, a DOM document, or a graph. The language's data adapter decides how to read it.

Choose an adapter

An adapter is the bridge between a language and the concrete shape of your data. Each provided language ships with a default:

Language	Default adapter	Expects
SQL	ObjectDataAdapter	array of objects; property = column
Cypher	GraphologyDataAdapter	a Graphology graph
XQuery	DomDataAdapter	a DOM document / XML tree

Pass a different one in the language's config when your data has another shape:

mainLang: SQL({ adapter: myAdapter });

SQL

By default SQL treats each element of a registered array as a row and each property as a column. See SQL Overview.

Cypher

The default Graphology adapter expects node labels and edge types under the gaLabelsOrType symbol key. See Cypher Overview. A plain labels property is not recognized, and label patterns then match nothing.

XQuery

DomDataAdapter targets DOM/XML. In the browser the global document is used automatically; in Node you pass a document from a DOM implementation such as jsdom:

mainLang: XQuery({ adapter: new DomDataAdapter(doc) });

Read a shape no adapter covers

Adapters are a first-class extension point: writing one is how you point a language at Map-backed rows, a different graph library, or a tree that is not a DOM. See Data Adapters.