apache/flink · error · IllegalArgumentException

Could not parse Avro schema string.

Error message

Could not parse Avro schema string.

What it means

AvroSchemaConverter.convertToTypeInfo(String, boolean) parses the supplied Avro schema string with Avro's Schema.Parser; SchemaParseException is wrapped in IllegalArgumentException('Could not parse Avro schema string.'). The string is not valid Avro schema JSON — syntax error, illegal type name, duplicate field, or malformed JSON.

Source

Thrown at flink-formats/flink-avro/src/main/java/org/apache/flink/formats/avro/typeutils/AvroSchemaConverter.java:125

    }

    /**
     * Converts an Avro schema string into a nested row structure with deterministic field order and
     * data types that are compatible with Flink's Table & SQL API.
     *
     * @param avroSchemaString Avro schema definition string
     * @param legacyTimestampMapping legacy mapping of timestamp types
     * @return type information matching the schema
     */
    @SuppressWarnings("unchecked")
    public static <T> TypeInformation<T> convertToTypeInfo(
            String avroSchemaString, boolean legacyTimestampMapping) {
        Preconditions.checkNotNull(avroSchemaString, "Avro schema must not be null.");
        final Schema schema;
        try {
            schema = new Schema.Parser().parse(avroSchemaString);
        } catch (SchemaParseException e) {
            throw new IllegalArgumentException("Could not parse Avro schema string.", e);
        }
        return (TypeInformation<T>) convertToTypeInfo(schema, legacyTimestampMapping);
    }

    private static TypeInformation<?> convertToTypeInfo(
            Schema schema, boolean legacyTimestampMapping) {
        switch (schema.getType()) {
            case RECORD:
                final List<Schema.Field> fields = schema.getFields();

                final TypeInformation<?>[] types = new TypeInformation<?>[fields.size()];
                final String[] names = new String[fields.size()];
                for (int i = 0; i < fields.size(); i++) {
                    final Schema.Field field = fields.get(i);
                    types[i] = convertToTypeInfo(field.schema(), legacyTimestampMapping);
                    names[i] = field.name();
                }
                return Types.ROW_NAMED(names, types);

View on GitHub (pinned to 2f3c205e92)

Solutions

1. Feed the exact schema content and validate it first: new Schema.Parser().parse(schemaString) in a unit test or at config load time to get the precise Avro error position.
2. Check for mangling: escaped quotes, HTML entities, truncation, or a path being read as literal text.
3. Use Avro tooling (avro-tools tojson/fromjson or an .avsc linter) to canonicalize the schema.

Example fix

// before
String schema = readConfig("schema.avsc"); // accidentally the PATH string
TypeInformation<?> ti = AvroSchemaConverter.convertToTypeInfo(schema);

// after
String schema = new String(Files.readAllBytes(Path.of("schema.avsc")), UTF_8);
new Schema.Parser().parse(schema); // fail fast with precise error
TypeInformation<?> ti = AvroSchemaConverter.convertToTypeInfo(schema);

try {
    new Schema.Parser().parse(avroSchemaString);
} catch (SchemaParseException e) {
    throw new IllegalArgumentException("Invalid Avro schema JSON: " + e.getMessage(), e);
}

try {
    return AvroSchemaConverter.convertToTypeInfo(avroSchemaString, legacy);
} catch (IllegalArgumentException e) {
    if (e.getCause() instanceof SchemaParseException) {
        // config-level bug: surface schema error, do not retry
        throw new ConfigurationException("Fix the .avsc: " + e.getCause().getMessage(), e);
    }
    throw e;
}

When it happens

Trigger: Calling convertToTypeInfo with a malformed .avsc string: trailing commas, wrong JSON, invalid names (fields starting with digits), duplicate record names, or accidentally passing a file path / classpath reference instead of the schema content.

Common situations: Loading schema strings from config/REST/CLI where quoting or escaping mangles JSON; passing a URL/path instead of content; hand-edited schemas with typos; copy-paste truncation.

Related errors

• Schema must be set when using Generic Record
• Unsupported type: {}
• Number of input splits has to be at least 1.
• Output file path may not be null.
• Delimiter must not be null

AI-assisted analysis of apache/flink@2f3c205e92 (2026-08-14). Data as JSON: /api/errors/af7784404112ab9a. Report an issue: GitHub.