apache/flink · error · RuntimeException

No parser available for type '{}'.

Error message

No parser available for type '{}'.

What it means

Thrown inside GenericCsvInputFormat.initializeSplit when FieldParser.getParserForType(fieldTypes[i]) returns null, meaning the configured field type has no registered CSV parser. Flink's CSV readers only support a fixed set of primitive/wrapper types (Byte, Short, Integer, Long, Float, Double, Boolean, String, BigDecimal, BigInteger, java.sql.Date, java.sql.Time, java.sql.Timestamp). This is a defensive runtime re-check; the same condition is normally caught earlier in setFieldTypesGeneric / setFieldsGeneric with an IllegalArgumentException, so reaching this RuntimeException means fieldTypes was populated without going through the validating setters.

Source

Thrown at flink-core/src/main/java/org/apache/flink/api/common/io/GenericCsvInputFormat.java:318

    }

    // --------------------------------------------------------------------------------------------
    //  Runtime methods
    // --------------------------------------------------------------------------------------------

    @Override
    protected void initializeSplit(FileInputSplit split, Long offset) throws IOException {
        super.initializeSplit(split, offset);

        // instantiate the parsers
        FieldParser<?>[] parsers = new FieldParser<?>[fieldTypes.length];

        for (int i = 0; i < fieldTypes.length; i++) {
            if (fieldTypes[i] != null) {
                Class<? extends FieldParser<?>> parserType =
                        FieldParser.getParserForType(fieldTypes[i]);
                if (parserType == null) {
                    throw new RuntimeException(
                            "No parser available for type '" + fieldTypes[i].getName() + "'.");
                }

                FieldParser<?> p = InstantiationUtil.instantiate(parserType, FieldParser.class);

                p.setCharset(getCharset());
                if (this.quotedStringParsing) {
                    if (p instanceof StringParser) {
                        ((StringParser) p).enableQuotedStringParsing(this.quoteCharacter);
                    } else if (p instanceof StringValueParser) {
                        ((StringValueParser) p).enableQuotedStringParsing(this.quoteCharacter);
                    }
                }

                parsers[i] = p;
            }
        }
        this.fieldParsers = parsers;

View on GitHub (pinned to 2f3c205e92)

Solutions

1. Restrict each configured field type to one of the supported parsers: Byte/byte, Short/short, Integer/int, Long/long, Float/float, Double/double, Boolean/boolean, String, BigDecimal, BigInteger, java.sql.Date, java.sql.Time, java.sql.Timestamp.
2. If you need a richer type, declare the field as String and convert it yourself in the map/function following the source.
3. Route field type setup through the validated setters (setFieldTypesGeneric / setFieldsGeneric) so an IllegalArgumentException is raised at job-construction time with a clearer message instead of at runtime on the cluster.
4. Check the unsupported type by calling FieldParser.getParserForType(type) in a unit test before submitting the job.

Example fix

// before
format.setFieldTypesGeneric(MyPojo.class, Integer.class);
// after
format.setFieldTypesGeneric(String.class, Integer.class);
// then map MyPojo from the String field yourself

import org.apache.flink.types.parser.FieldParser;

private static void assertAllTypesParsable(Class<?>... fieldTypes) {
    for (Class<?> t : fieldTypes) {
        if (t != null && FieldParser.getParserForType(t) == null) {
            throw new IllegalArgumentException(
                "No CSV parser for type " + t.getName()
                + ". Supported: primitives/wrappers, String, BigDecimal, BigInteger,"
                + " java.sql.Date/Time/Timestamp.");
        }
    }
}
// call in a unit test or job setup before opening the format

// Restrict the field-type vocabulary at the API boundary
static final Set<Class<?>> CSV_SUPPORTED = Set.of(
    Byte.class, Short.class, Integer.class, Long.class, Float.class, Double.class,
    Boolean.class, String.class, BigDecimal.class, BigInteger.class,
    java.sql.Date.class, java.sql.Time.class, java.sql.Timestamp.class);

static Class<?> csvType(Class<?> t) {
    if (!CSV_SUPPORTED.contains(t)) throw new IllegalArgumentException("Unsupported CSV type: " + t);
    return t;
}

// Wrap format open in a try and surface a clear configuration error
try {
    format.open(split);
} catch (RuntimeException e) {
    if (e.getMessage() != null && e.getMessage().startsWith("No parser available for type")) {
        throw new IllegalArgumentException(
            "A configured CSV field type has no parser. Use a supported type.", e);
    }
    throw e;
}

When it happens

Trigger: A subclass of GenericCsvInputFormat (or RowCsvInputFormat) assigns this.fieldTypes directly (or via a custom path) with an unsupported Class such as a POJO, Tuple, Map, List, java.util.Date, java.time.LocalDate, or an enum. Then JobManager/TaskManager opens the split and initializeSplit runs getParserForType which returns null.

Common situations: Switching from java.util.Date to java.time types after a Java 8 migration; passing a custom value class thinking the CSV reader will call a constructor/fromString method; using a Tuple type field instead of its component types; upgrading Flink where a previously-tolerated type is no longer auto-converted; copy-pasting a Class<?> array from a different reader.

Related errors

• The type '{}' is not supported for the CSV input format.
• Delimiter must not be null
• Field types must not be null.
• Field indices must not be smaller than zero.
• Missing type for included field {}.

AI-assisted analysis of apache/flink@2f3c205e92 (2026-08-14). Data as JSON: /api/errors/6a56cb419c3371c9. Report an issue: GitHub.