prestodb/presto · error · PrestoException

NOT_FOUND

NOT_FOUND

Error message

Materialized view not found:

What it means

An Iceberg materialized view operation attempted to read the backing metastore table but the metastore threw TableNotFoundException. Presto translates that into NOT_FOUND with 'Materialized view not found: <name>', meaning no table exists under that schema/name in the metastore.

Source

Thrown at presto-iceberg/src/main/java/com/facebook/presto/iceberg/IcebergHiveMetadata.java:839

        catch (TableAlreadyExistsException e) {
            throw new PrestoException(ALREADY_EXISTS, "Materialized view already exists: " + viewName);
        }
    }

    @Override
    protected void dropIcebergView(ConnectorSession session, SchemaTableName schemaTableName)
    {
        MetastoreContext metastoreContext = getMetastoreContext(session);

        try {
            metastore.dropTable(
                    metastoreContext,
                    schemaTableName.getSchemaName(),
                    schemaTableName.getTableName(),
                    true);
        }
        catch (TableNotFoundException e) {
            throw new PrestoException(NOT_FOUND, "Materialized view not found: " + schemaTableName);
        }
    }

    @Override
    protected void updateIcebergViewProperties(
            ConnectorSession session,
            SchemaTableName viewName,
            Map<String, String> properties)
    {
        MetastoreContext metastoreContext = getMetastoreContext(session);
        try (HiveMetastoreLock ignored = HiveMetastoreLock.acquire(
                metastore,
                metastoreContext,
                hiveTableOperationsConfig.getLockingEnabled(),
                viewName.getSchemaName(),
                viewName.getTableName())) {
            Optional<Table> existingTable = getHiveTable(session, viewName);
            if (!existingTable.isPresent() || !isIcebergMaterializedView(existingTable.get())) {

View on GitHub (pinned to 55bb57d202)

Solutions

  1. Verify the fully qualified name with SHOW TABLES FROM <schema> (MVs appear as tables)
  2. Create the materialized view if it never existed: CREATE MATERIALIZED VIEW ...
  3. Check you are querying the intended catalog (iceberg vs hive connector namespace)
  4. If the backing table was deleted accidentally, restore it from the metastore/backup

Example fix

// before
REFRESH MATERIALIZED VIEW sales.daily_mv; -- NOT_FOUND
// after
SHOW TABLES FROM sales; -- confirm name
CREATE MATERIALIZED VIEW sales.daily_mv AS SELECT ...; -- or fix the name
Defensive patterns

Strategy: try-catch

Validate before calling

-- Verify the MV exists before operating on it:
SHOW TABLES FROM sales LIKE 'daily_mv';

Try / catch

try {
    stmt.execute("REFRESH MATERIALIZED VIEW sales.daily_mv");
} catch (SQLException e) {
    if (e.getMessage() != null && e.getMessage().contains("not found")) {
        // create the MV or fix the catalog/schema/name, then retry
    } else { throw e; }
}

Prevention

When it happens

Trigger: REFRESH/DROP or other MV operations referencing a materialized view name that was never created or has been dropped; querying after the backing table was removed out-of-band via the metastore.

Common situations: Typos in schema or view name; wrong catalog prefix; another team/job dropped the backing table directly in Hive; environment drift between dev and prod names.

Understand the failure class

Background: NOT_FOUND error code: why tRPC, Harbor, Nacos and other libraries return 404 "not found" errors for resources that may still exist — this error's family across 11 libraries.

Related errors

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/c54644979bb404bf. Report an issue: GitHub.