Arbitrarily large sequences of values pose a problem for JavaScript Object Notation (JSON) [RFC7159] that is well explained in the motivation for JSON Text Sequences [RFC7464]. The GeoJSON format [RFC7946] faces the same kind of problem. Geographic datasets often run to the tens of thousands or millions of features. The problem is often amplified by the presence of large arrays of coordinates for each of the features.
This document describes a specialization of JSON Text Sequences. A GeoJSON Text Sequence is a document of arbitrarily large size containing one or more GeoJSON objects, e.g., multiple GeoJSON texts that can be produced and parsed incrementally, and not only a single GeoJSON FeatureCollection, Feature, or Geometry.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
Defined in prose, following [RFC7464]: a GeoJSON text sequence is any number of GeoJSON [RFC7946] texts, each encoded in UTF-8 [RFC3629], each preceded by one ASCII [RFC20] RS character, and each followed by a line feed (LF).
The GeoJSON Text Sequence Format conforms to all the rules of [RFC7464] and adds the following constraint: each JSON text MUST contain a single GeoJSON object as defined in [RFC7946].
Heterogeneous sequences containing a mix of GeoJSON Geometry, Feature, and FeatureCollection objects are permitted. How producers and parsers of GeoJSON text sequences communicate rules for allowed GeoJSON types in exchanged sequences is not specified in this document.