SeriesSplitter Troubleshooting - Teradata Vantage

Problem: Invoking a function using SeriesSplitter does not improve execution time.

Before trying workaround, verify that the data is skewed and that the function that uses SeriesSplitter does not exploit full parallelism. If the data is not skewed and the function exploits full parallelism, SeriesSplitter cannot improve its execution time.

Workaround:

• Invoke SeriesSplitter and the subsequent function in separate SQL calls. For example, in this SQL call, company1_stock1_split is the output from SeriesSplitter Example: Partition Splitter:

  SELECT * FROM Interpolator (
    ON company1_stock1_split AS InputTable
    PARTITION BY id ORDER BY "period"
    USING
    TimeColumn ('period')
    TimeInterval (86400)
    InterpolationType ('linear')
    ValueColumns ('stockprice')
    Accumulate ('id')
    DuplicateRowsCount (2)
  ) AS dt;

  The preceding SQL call is likely to perform better than the following call, which uses SeriesSplitter in the ON clause of the subsequent function.

  SELECT * FROM Interpolator (
    ON (
      SELECT * FROM SeriesSplitter (
        ON company1_stock1 AS InputTable
        USING
        PartitionByColumns ('id')
        OrderByColumns ('period')
        SplitCount (50)
        Accumulate ('stockprice')
        ReturnStatsTable ('false')
      ) AS dt1
    ) AS input_table PARTITION BY id ORDER BY period
    USING
    TimeColumn ('period')
    TimeInterval (86400)
    InterpolationType ('linear')
    ValueColumns ('stockprice')
    Accumulate ('id')
    DuplicateRowsCount (2)
  ) AS dt2;

• Adjust these syntax elements as follows:

  Syntax Element	Value
  DuplicateRowsCount	As low as possible.
  SplitCount	A smaller multiple (for example, 1) of the number of vworkers in the cluster.
  RowsPerSplit	As high as possible (you want the resulting number of splits to be a smaller multiple of the number of vworkers in the cluster).
  Accumulate	Specify as few columns as possible.
  DuplicateColumn	Omit this syntax element.
  PartialSplitID	'true'
  ReturnStatsTable	'true'