Package org.apache.avro

Class JsonSchemaParser

java.lang.Object

org.apache.avro.JsonSchemaParser

All Implemented Interfaces:

FormattedSchemaParser

public class JsonSchemaParser
extends Object
implements FormattedSchemaParser

Schema parser for JSON formatted schemata. This initial implementation simply delegates to the Schema.Parser class, though it should be refactored out of there.

Note: this class is intentionally not available via the Java ServiceLoader, as its use is hardcoded as fallback when no service exists. This enables users to reliably override the standard JSON parser as well.

Constructor Summary

Constructors

Constructor	Description
JsonSchemaParser()

Method Summary

Modifier and Type	Method	Description
Schema	parse(ParseContext parseContext, URI baseUri, CharSequence formattedSchema)	Parse schema definitions from a text based source.
static Schema	parseInternal(String... fragments)	Parse a schema written in the internal (JSON) format without any validations.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

JsonSchemaParser

public JsonSchemaParser()

Method Details

parseInternal

public static Schema parseInternal(String... fragments)

Parse a schema written in the internal (JSON) format without any validations.

Using this method is only safe if used to parse a write schema (i.e., a schema used to read Avro data). Other usages, for example by generated Avro code, can cause interoperability problems.

Use with care and sufficient testing!

Parameters:

fragments - one or more strings making up the schema (some schemata exceed the compiler limits)

Returns:

the parsed schema

parse

public Schema parse(ParseContext parseContext,
 URI baseUri,
 CharSequence formattedSchema)
             throws IOException,
SchemaParseException

Description copied from interface: FormattedSchemaParser

Parse schema definitions from a text based source.

Notes for implementers:

• Schema definitions are expected not to be in the format the parser expects. So when the input clearly doesn't make sense (e.g., reading "/**" when expecting JSON), it is a good idea not to do anything (especially calling methods on the @code ParseContext}).
• The parameter parseContext is not thread-safe.
• All named schema definitions that are parsed should be added to the provided ParseContext.
• Optionally, you may return a "main" schema. Some schema definitions have one, for example the schema defined by the root of the JSON document in a standard schema definition. If unsure, return null.
• If parsing fails, throw a SchemaParseException. This will let the parsing process recover and continue.
• Throwing anything other than a SchemaParseException will abort the parsing process, so reserve that for rethrowing exceptions.

Specified by:

parse in interface FormattedSchemaParser

Parameters:

parseContext - the current parse context: all named schemata that are parsed should be added here, otherwise resolving schemata can fail; contains all previously known types

baseUri - the base location of the schema, or null if not known

formattedSchema - the text of the schema definition(s) to parse

Returns:

the main schema, if any

Throws:

IOException - when the schema cannot be read

SchemaParseException - when the schema cannot be parsed