apache/iceberg · error · UnsupportedOperationException

Operation manageSnapshots is not supported after the table…

Error message

Operation manageSnapshots is not supported after the table is serialized

What it means

SerializableTable is a read-only serializable copy of a table for distribution across cluster nodes; it cannot commit changes. manageSnapshots() (cherrypick/rollback of snapshots) is a mutating operation, so the implementation always throws this UnsupportedOperationException. Snapshot management must be performed on a table loaded directly from its catalog.

Solutions

Reload the table from its catalog (catalog.loadTable(location)) on the node that performs the snapshot operation, then call manageSnapshots().
Perform the operation in the driver where the original catalog-backed Table is available, and only ship SerializableTable for read/scan work.
Add an instanceof check before invoking snapshot-management APIs and fail fast with a clear message or route to a catalog-loaded table.

Example fix

// before
table.manageSnapshots().rollbackTo(snapshotId); // table is SerializableTable

// after
Table fresh = catalog.loadTable(tableLocation);
fresh.manageSnapshots().rollbackTo(snapshotId);

Defensive patterns

Strategy: type-guard

Validate before calling

if (table instanceof org.apache.iceberg.SerializableTable) {
  table = catalog.loadTable(table.location()); // reload before managing snapshots
}

Type guard

boolean supportsSnapshotManagement(Table t) {
  return !(t instanceof org.apache.iceberg.SerializableTable);
}

Try / catch

try {
  table.manageSnapshots().rollbackTo(snapshotId);
} catch (UnsupportedOperationException e) {
  catalog.loadTable(table.location()).manageSnapshots().rollbackTo(snapshotId);
}

Prevention

Perform rollback/cherrypick in the driver where the table came from a Catalog.
Treat any Table that crossed a serialization boundary as read-only.
Prefer engine action APIs (e.g. Spark rollback procedures) over raw Table APIs on distributed references.

When it happens

Trigger: Calling table.manageSnapshots() (e.g. rollback, cherrypick, createBranch) on a Table that is a SerializableTable — after Java serialization to an executor or via SerializableTable.copyOf().

Common situations: Executing rollback/cherrypick logic in a distributed task where the table arrived serialized inside the closure; using a deserialized table in custom engine integrations; mistakenly treating a broadcast table as a normal catalog table.

Understand the failure class

Background: UnsupportedOperationException and "is not supported" errors: when a library deliberately refuses a call — this error's family across 30 libraries.

Related errors

AI-assisted analysis of apache/iceberg@86d9c8fc54 (2026-09-12). Data as JSON: /api/errors/f38108f596c00267. Report an issue: GitHub.

Appendix: source

Thrown at core/src/main/java/org/apache/iceberg/SerializableTable.java:427

  @Override
  public UpdateStatistics updateStatistics() {
    throw new UnsupportedOperationException(errorMsg("updateStatistics"));
  }

  @Override
  public UpdatePartitionStatistics updatePartitionStatistics() {
    throw new UnsupportedOperationException(errorMsg("updatePartitionStatistics"));
  }

  @Override
  public ExpireSnapshots expireSnapshots() {
    throw new UnsupportedOperationException(errorMsg("expireSnapshots"));
  }

  @Override
  public ManageSnapshots manageSnapshots() {
    throw new UnsupportedOperationException(errorMsg("manageSnapshots"));
  }

  @Override
  public Transaction newTransaction() {
    throw new UnsupportedOperationException(errorMsg("newTransaction"));
  }

  @Override
  public StaticTableOperations operations() {
    return (StaticTableOperations) ((BaseTable) lazyTable()).operations();
  }

  private String errorMsg(String operation) {
    return String.format("Operation %s is not supported after the table is serialized", operation);
  }

  public static class SerializableMetadataTable extends SerializableTable {
    private final MetadataTableType type;

View on GitHub (pinned to 86d9c8fc54)