prestodb/presto · error · PrestoException

NOT_SUPPORTED

NOT_SUPPORTED

Error message

Catalog "%s" cannot be used as a partitioning provider: %s

What it means

AddExchanges delegates hash partitioning to a catalog's partitioning provider via getPartitioningHandleForExchange. If the provider catalog throws NOT_SUPPORTED (it cannot supply that partitioning), AddExchanges wraps it in a clearer NOT_SUPPORTED error naming the catalog.

Source

Thrown at presto-main-base/src/main/java/com/facebook/presto/sql/planner/optimizations/AddExchanges.java:1754

        }

        private Partitioning createPartitioning(Collection<VariableReferenceExpression> partitioningColumns)
        {
            // TODO: Use SystemTablesMetadata instead of introducing a special case
            if (GlobalSystemConnector.NAME.equals(partitioningProviderCatalog)) {
                return Partitioning.create(FIXED_HASH_DISTRIBUTION, ImmutableList.copyOf(partitioningColumns));
            }

            List<Type> partitioningTypes = partitioningColumns.stream()
                    .map(VariableReferenceExpression::getType)
                    .collect(toImmutableList());
            try {
                PartitioningHandle partitioningHandle = metadata.getPartitioningHandleForExchange(session, partitioningProviderCatalog, hashPartitionCount, partitioningTypes);
                return Partitioning.create(partitioningHandle, partitioningColumns);
            }
            catch (PrestoException e) {
                if (e.getErrorCode().equals(NOT_SUPPORTED.toErrorCode())) {
                    throw new PrestoException(
                            NOT_SUPPORTED,
                            format(
                                    "Catalog \"%s\" cannot be used as a partitioning provider: %s",
                                    partitioningProviderCatalog,
                                    e.getMessage()),
                            e);
                }
                throw e;
            }
        }

        // TODO: refactor this method into ExchangeNode#partitionedExchange once
        //   materialized exchange is supported for all nodes.
        private Scope selectExchangeScopeForPartitionedRemoteExchange(PlanNode exchangeSource, boolean nullsAndAnyReplicated)
        {
            if (nullsAndAnyReplicated || exchangeSource.getOutputVariables().isEmpty()) {
                // materialized remote exchange is not supported when
                //  * replicateNullsAndAny is needed

View on GitHub (pinned to 55bb57d202)

Solutions

1. Remove or fix the 'partitioning_provider_catalog' session property / connector config to point at a catalog supporting the partitioning
2. Avoid cross-catalog joins that force partitioning through an unsupported connector, or co-locate data in one catalog
3. Check the wrapped cause (e.getMessage()) for the connector-specific limitation and adjust partition count or types

Example fix

// before
SET SESSION partitioning_provider_catalog = 'iceberg'; -- unsupported partitioning
SELECT ... JOIN ...;
// after
RESET SESSION partitioning_provider_catalog;
SELECT ... JOIN ...;

// confirm connector supports exchange partitioning before cross-catalog joins
SHOW SESSION LIKE 'partitioning_provider_catalog';

null

try {
    return query(sql);
} catch (PrestoException e) {
    if (e.getErrorCode() == NOT_SUPPORTED.toErrorCode() && e.getMessage().contains("partitioning provider")) {
        resetSessionProperty("partitioning_provider_catalog");
        return retryQuery(sql);
    }
    throw e;
}

When it happens

Trigger: A query's join/aggregation distribution requires a partitioning handle from a connector catalog that does not support the requested partitioning type or partition count.

Common situations: Querying tables across catalogs where one connector lacks partitioning support; partitioning_provider_catalog misconfigured; hash partition count unsupported by the connector.

Understand the failure class

Background: Presto NOT_SUPPORTED error: what "not supported" means and how to fix it — this error's family across 3 libraries.

Related errors

• GENERIC_INTERNAL_ERROR
• ARROW_INTERNAL_ERROR
• NOT_SUPPORTED
• NOT_FOUND
• JDBC_ERROR

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/ff7a6a69ead1c060. Report an issue: GitHub.