Packages

package backfill

Type Members

  1. trait BackfillBatch extends DeltaLogging
  2. class BackfillBatchIterator[BackfillBatchType] extends Iterator[BackfillBatchType] with Closeable

    Construct a lazily evaluated iterator for getting BackfillBatchType.

    Construct a lazily evaluated iterator for getting BackfillBatchType. Callers should close this iterator to avoid resource leak.

  3. case class BackfillBatchStats(parentTransactionId: String, transactionId: String, batchId: Int, initialNumFiles: Long, totalExecutionTimeInMs: Long, wasSuccessful: Boolean) extends Product with Serializable

    Metrics for each BackfillBatch.

    Metrics for each BackfillBatch.

    parentTransactionId

    The transaction id associated with the parent command.

    transactionId

    The transaction id used in this batch.

    batchId

    An integer identifier of the batch within a parent BackfillCommand.

    initialNumFiles

    The number of files in BackfillBatch prior to conflict resolution.

    totalExecutionTimeInMs

    The total execution time in milliseconds.

    wasSuccessful

    Boolean indicating whether the batch was successfully committed.

  4. trait BackfillCommand extends LogicalPlan with LeafRunnableCommand with DeltaCommand

    This command will lazily materialize AllFiles and split them into multiple backfill commits if the number of files exceeds the threshold set by DeltaSQLConf.DELTA_BACKFILL_MAX_NUM_FILES_PER_COMMIT.

  5. case class BackfillCommandStats(transactionId: String, nameOfTriggeringOperation: String, maxNumBatchesInParallel: Int, totalExecutionTimeMs: Long = 0, numSuccessfulBatches: Int = 0, numFailedBatches: Int = 0, wasSuccessful: Boolean = false) extends Product with Serializable

    Metrics for the BackfillCommand.

    Metrics for the BackfillCommand.

    maxNumBatchesInParallel

    The maximum number of batches that could have run in parallel.

    totalExecutionTimeMs

    The total execution time in milliseconds.

    numSuccessfulBatches

    The number of BackfillBatch's that was successfully committed.

    numFailedBatches

    The number of BackfillBatch's that failed.

    wasSuccessful

    Boolean indicating whether this BackfillCommand didn't have any error.

  6. trait BackfillExecutor extends DeltaLogging
  7. case class RowTrackingBackfillBatch(filesInBatch: Seq[AddFile]) extends BackfillBatch with Product with Serializable
  8. case class RowTrackingBackfillCommand(deltaLog: DeltaLog, nameOfTriggeringOperation: String, catalogTable: Option[CatalogTable]) extends LogicalPlan with BackfillCommand with Product with Serializable

    This command re-commits all AddFiles in the current snapshot that do not have a base row IDs.

    This command re-commits all AddFiles in the current snapshot that do not have a base row IDs. After the backfill command finishes, the snapshot has row IDs for all files. All commits afterwards must include row IDs.

    First, we will add the table feature support, if necessary. Then, the command will lazily materialize AllFiles and split them into multiple backfill commits if the number of files exceeds the threshold set by DeltaSQLConf.DELTA_BACKFILL_MAX_NUM_FILES_PER_COMMIT.

    Note: We expect Backfill to be called before the table property is set. Furthermore, we do not set the table property DeltaConfigs.ROW_TRACKING_ENABLED as part of backfill. The metadata update needs to be handled by the caller.

  9. class RowTrackingBackfillExecutor extends BackfillExecutor

Ungrouped