Abort all the updates that have been made to the store. Implementations should ensure that no more updates (puts, removes) can be after an abort in order to avoid incorrect usage.

Commit all the updates that have been made to the store, and return the new version. Implementations should ensure that no more updates (puts, removes) can be after a commit in order to avoid incorrect usage.

Create column family with given name, if absent.

Get the current value of a non-null key.

Return information on recently generated checkpoints The information should only be usable when checkpoint format version 2 is used and underlying state store supports it. If it is not the case, the method can return a dummy result. The result eventually won't be sent to the driver, but not all the stateful operator is able to figure out whether the function should be called to now. They would anyway call it and pass it to StatefulOperator.setStateStoreCheckpointInfo(), where it will be ignored.

Unique identifier of the store

Return an iterator containing all the key-value pairs in the StateStore. Implementations must ensure that updates (puts, removes) can be made while iterating over this iterator.

It is not required for implementations to ensure the iterator reflects all updates being performed after initialization of the iterator. Callers should perform all updates before calling this method if all updates should be visible in the returned iterator.

Merges the provided value with existing values of a non-null key. If a existing value does not exist, this operation behaves as StateStore.put().

It is expected to throw exception if Spark calls this method without setting multipleValuesPerKey as true for the column family.

Return an iterator containing all the key-value pairs which are matched with the given prefix key.

The operator will provide numColsPrefixKey greater than 0 in StateStoreProvider.init method if the operator needs to leverage the "prefix scan" feature. The schema of the prefix key should be same with the leftmost numColsPrefixKey columns of the key schema.

It is expected to throw exception if Spark calls this method without setting numColsPrefixKey to the greater than 0.

Put a new non-null value for a non-null key. Implementations must be aware that the UnsafeRows in the params can be reused, and must make copies of the data as needed for persistence.

Provides an iterator containing all values of a non-null key. If key does not exist, an empty iterator is returned. Implementations should make sure to return an empty iterator if the key does not exist.

It is expected to throw exception if Spark calls this method without setting multipleValuesPerKey as true for the column family.

Version of the data in this store before committing updates.