Json Extraction - Usage
Table of Contents
The main entry point is
Learn() method, which returns a
Program’s key method is
Run() that executes the program on an input Json to obtain the extracted output.
Each program also has a
Schema property that defines the structure of the extracted data.
Other important methods are
Deserialize() to serialize and deserialize
Extraction.Json, one needs to reference
The Sample Project illustrates our API usage.
By default, Extraction.Json learns a join program in which inner arrays are joined with other fields. As a result, an outer object in the input Json can be flattened into several rows in the output table.
The below snippet illustrates a learning session to generate such program from the input
string jsonText = ... var session = new Session(); session.Constraints.Add(new FlattenDocument(jsonText)); Program program = session.Learn();
Clients may add
NoJoinInnerArrays constraint to the session to learn
non-join programs, as illustrated in the following snippet:
var noJoinSession = new Session(); noJoinSession.Constraints.Add(new FlattenDocument(jsonText), new NoJoinInnerArrays()); Program noJoinProgram = noJoinSession.Learn();
The Introduction page has more discussion on this topic.
Serializing/Deserializing a Program
Extraction.Json.Program.Serialize() method serializes the learned program to a string.
Extraction.Json.Loader.Instance.Load() method deserializes the program text to a program.
// program was learned previously string progText = program.Serialize(); Program loadProg = Loader.Instance.Load(progText);
Executing a Program
Given an input Json, a program can generate a hierarchical tree or a flattened table. If the program is a join program, the table is flattened either using outer join (default) or inner join semantics.
Generating a Tree
Use this method to obtain a hierarchical tree of the input document.
// program was learned previously ITreeOutput<JsonRegion> tree = program.Run(jsonText);
Generating a Table
Supply the desired join semantics to the
RunTable() method as follows:
// program was learned previously IEnumerable<TableRow<JsonRegion>> outerJoinTable = program.RunTable(jsonText, TreeToTableSemantics.OuterJoin); IEnumerable<TableRow<JsonRegion>> innerJoinTable = program.RunTable(jsonText, TreeToTableSemantics.InnerJoin);