JSONiqlike

JSONiqlike is a proposed query language designed for declarative access and transformation of JSON-like data structures. It combines elements of SQL with JSON-oriented path expressions to enable querying nested objects and arrays without writing procedural code.

Its design emphasizes readability, schema-free operation, and compatibility with common JSON encodings. Key capabilities include extracting fields, flattening arrays, applying filters, performing aggregations, and joining data across multiple JSON sources.

Syntax and evaluation model: JSONiqlike uses a SELECT-like clause to specify output fields, FROM to name data sources (files, streams, or databases), and WHERE for predicates. Path expressions navigate the JSON document, supporting dot notation, bracket notation, and array wildcards, while operators handle comparison, existence, and logical composition. Some proposals add UNNEST or FLATTEN to materialize array elements for downstream processing.

Example (in prose): SELECT customer.name, customer.address.city FROM customers WHERE customer.membership = 'gold'. The language aims to be amenable to both ad hoc analysis and data integration pipelines.

Relationship to related technologies: JSONiqlike draws on JSONiq, SQL, and JQ, but seeks to offer SQL-like readability with JSON-centric semantics. It is not a standard; as of now, there are no widely adopted implementations or certified specifications, and variants exist only in research or experimental projects.