prestodb/presto · error · PrestoException

NOT_SUPPORTED

NOT_SUPPORTED

Error message

Inserting into bucketed tables with skew is not supported. %s

What it means

checkWritable refuses writes to Hive tables/partitions that are both bucketed and skewed. Presto's bucketing writer cannot preserve the skewed-bucket layout the table declares, so writing would corrupt bucketing invariants. The exception carries the table/partition description for diagnosis.

Source

Thrown at presto-hive/src/main/java/com/facebook/presto/hive/HiveWriteUtils.java:405

            Map<String, String> parameters,
            Storage storage)
    {
        String tablePartitionDescription = "Table '" + tableName + "'";
        if (partitionName.isPresent()) {
            tablePartitionDescription += " partition '" + partitionName.get() + "'";
        }

        // verify online
        verifyOnline(tableName, partitionName, protectMode, parameters);

        // verify not read only
        if (protectMode.readOnly) {
            throw new HiveReadOnlyException(tableName, partitionName);
        }

        // verify skew info
        if (storage.isSkewed()) {
            throw new PrestoException(NOT_SUPPORTED, format("Inserting into bucketed tables with skew is not supported. %s", tablePartitionDescription));
        }
    }

    public static Path getTableDefaultLocation(ConnectorSession session, SemiTransactionalHiveMetastore metastore, HdfsEnvironment hdfsEnvironment, String schemaName, String tableName)
    {
        MetastoreContext metastoreContext = new MetastoreContext(
                session.getIdentity(),
                session.getQueryId(),
                session.getClientInfo(),
                session.getClientTags(),
                session.getSource(),
                getMetastoreHeaders(session),
                isUserDefinedTypeEncodingEnabled(session),
                metastore.getColumnConverterProvider(),
                session.getWarningCollector(),
                session.getRuntimeStats());
        Optional<String> location = getDatabase(session.getIdentity(), metastoreContext, metastore, schemaName).getLocation();
        if (!location.isPresent() || location.get().isEmpty()) {

View on GitHub (pinned to 55bb57d202)

Solutions

1. Remove the skew specification: recreate the table without SKEWED BY, copy data, and swap names
2. Write to a different (non-skewed) staging table from Presto, then load into the skewed table via Hive itself
3. If skew metadata is stale, drop the skew info in the metastore (ALTER TABLE ... NOT SKEWED / update table properties) if it no longer reflects reality

Example fix

// before
CREATE TABLE t (k int) CLUSTERED BY (k) INTO 4 BUCKETS SKEWED BY (k) ON (1,2); -- insert fails
// after
CREATE TABLE t (k int) CLUSTERED BY (k) INTO 4 BUCKETS;

Table t = metastore.getTable(schema, table);
if (t.getStorage().isSkewed() && t.getStorage().getBucketProperty().isPresent()) {
    throw new IllegalStateException("Target table is bucketed+skewed; not writable from Presto");
}

try {
    insertInto(skewedTable);
} catch (PrestoException e) {
    if (NOT_SUPPORTED.toErrorCode().equals(e.getErrorCode()) && e.getMessage().contains("bucketed tables with skew")) {
        // fall back to a staging non-skewed table
    } else throw e;
}

When it happens

Trigger: INSERT (or CTAS writing into an existing table) whose target HiveTableHandle/storage has isSkewed() true — i.e. the table was created with SKEWED BY and is also bucketed (CLUSTERED BY ... INTO N BUCKETS).

Common situations: Inserting into legacy Hive tables defined with SKEWED BY clause created outside Presto; schema migration bringing skewed bucketed tables into a Presto write path.

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

• NOT_SUPPORTED
• unknown java type
• HIVE_INVALID_BUCKET_FILES
• HIVE_INVALID_METADATA
• NOT_SUPPORTED

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/3eb52fd8da0b0976. Report an issue: GitHub.