prestodb/presto · error · IllegalArgumentException

unknown data format '%s' for column '%s'

Error message

unknown data format '%s' for column '%s'

What it means

HashRedisRowDecoderFactory.chooseFieldDecoder selects a field decoder based on the column's dataFormat. For hash decoders only the default (null dataFormat) and 'iso8601' for BIGINT (long) columns are supported; any other dataFormat string throws IllegalArgumentException naming the format and column.

Source

Thrown at presto-redis/src/main/java/com/facebook/presto/redis/decoder/hash/HashRedisRowDecoderFactory.java:55

        return new HashRedisRowDecoder(chooseFieldDecoders(columns));
    }

    private Map<DecoderColumnHandle, RedisFieldDecoder<String>> chooseFieldDecoders(Set<DecoderColumnHandle> columns)
    {
        return columns.stream()
                .collect(ImmutableMap.toImmutableMap(identity(), this::chooseFieldDecoder));
    }

    private RedisFieldDecoder<String> chooseFieldDecoder(DecoderColumnHandle column)
    {
        checkArgument(!column.isInternal(), "unexpected internal column '%s'", column.getName());
        if (column.getDataFormat() == null) {
            return new HashRedisFieldDecoder();
        }
        if (column.getType().getJavaType() == long.class && "iso8601".equals(column.getDataFormat())) {
            return new ISO8601HashRedisFieldDecoder();
        }
        throw new IllegalArgumentException(format("unknown data format '%s' for column '%s'", column.getDataFormat(), column.getName()));
    }
}

View on GitHub (pinned to 55bb57d202)

Solutions

1. Remove the unsupported dataFormat attribute so the default HashRedisFieldDecoder is used
2. Use dataFormat='iso8601' only on BIGINT columns
3. If a timestamp string is needed, store/declare as VARCHAR without iso8601 and parse in SQL with date_parse

Example fix

// before
event_time TIMESTAMP WITH (dataFormat='iso8601') -- hash decoder requires BIGINT
// after
event_time BIGINT WITH (dataFormat='iso8601')

String df = column.getDataFormat();
if (df != null && !("iso8601".equals(df) && column.getType() == BIGINT)) {
    throw new IllegalArgumentException("unsupported dataFormat for hash decoder: " + df);
}

try { Decoder d = HashRedisRowDecoderFactory.create(...); } catch (IllegalArgumentException e) { log.error("bad column definition: " + e.getMessage()); throw e; }

When it happens

Trigger: A Redis hash-table column declares an unsupported dataFormat value (e.g. 'epochmillis', 'bytes', 'json') or uses 'iso8601' on a non-BIGINT (e.g. VARCHAR) column.

Common situations: Copying decoder options from other connectors (Kafka supports more formats) into Redis hash DDL; typo in 'iso8601'; applying iso8601 to a VARCHAR timestamp column.

Related errors

• ALREADY_EXISTS
• INVALID_VIEW
• ACCUMULO_TABLE_EXISTS
• UNEXPECTED_ACCUMULO_ERROR
• ACCUMULO_TABLE_EXISTS

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/8daa8afd09f4d1e1. Report an issue: GitHub.