Window functions

Teiid provides ANSI SQL 2003 window functions. A window function allows an aggregate function to be applied to a subset of the result set, without the need for a GROUP BY clause. A window function is similar to an aggregate function, but requires the use of an OVER clause or window specification.

Usage:
  aggregate [FILTER (WHERE ...)] OVER ( [partition] [ORDER BY ...] [frame] )
| FIRST_VALUE(val) OVER ( [partition] [ORDER BY ...] [frame] )
| LAST_VALUE(val) OVER ( [partition] [ORDER BY ...] [frame] )
| analytical OVER ( [partition] [ORDER BY ...] )

partition := PARTITION BY expression [, expression]*

frame := range_or_rows extent

range_or_rows := RANGE | ROWS

extent :=
    frameBound
  | BETWEEN frameBound AND frameBound

frameBound :=
    UNBOUNDED PRECEDING
  | UNBOUNDED FOLLOWING
  | n PRECEDING
  | n FOLLOWING
  | CURRENT ROW

In the preceding syntax, aggregate can refer to any aggregate function. Keywords exist for the following analytical functions ROW_NUMBER, RANK, DENSE_RANK, PERCENT_RANK, CUME_DIST. There are also the FIRST_VALUE, LAST_VALUE, LEAD, LAG, NTH_VALUE, and NTILE analytical functions. For more information, see Analytical functions definitions.

Syntax rules

  • Window functions can only appear in the SELECT and ORDER BY clauses of a query expression.

  • Window functions cannot be nested in one another.

  • Partitioning and order by expressions cannot contain subqueries or outer references.

  • An aggregate ORDER BY clause cannot be used when windowed.

  • The window specification ORDER BY clause cannot reference alias names or use positional ordering.

  • Windowed aggregates may not use DISTINCT if the window specification is ordered.

  • Analytical value functions may not use DISTINCT and require the use of an ordering in the window specification.

  • RANGE or ROWS requires the ORDER BY clause to be specified. The default frame if not specified is RANGE UNBOUNDED PRECEDING. If no end is specified the default is CURRENT ROW. No combination of start and end is allowed such that the end is before the start - for example UNBOUNDED FOLLOWING is not allow as a start nor is UNBOUNDED PRECEDING allowed as an end.

  • RANGE cannot be used n PRECEDING or n FOLLOWING

Analytical function definitions
Ranking functions

  • RANK() – Assigns a number to each unique ordering value within each partition starting at 1, such that the next rank is equal to the count of prior rows.

  • DENSE_RANK() – Assigns a number to each unique ordering value within each partition starting at 1, such that the next rank is sequential.

  • PERCENT_RANK() – Computed as (RANK - 1) / ( RC - 1) where RC is the total row count of the partition.

  • CUME_DIST() – Computed as the PR / RC where PR is the rank of the row including peers and RC is the total row count of the partition.

    By default all values are integers - an exception will be thrown if a larger value is needed. Use the system org.teiid.longRanks to have RANK, DENSE_RANK, and ROW_NUMBER return long values instead.

Value functions

  • FIRST_VALUE(val) – Return the first value in the window frame with the given ordering.

  • LAST_VALUE(val) – Return the last observed value in the window frame with the given ordering.

  • LEAD(val [, offset [, default]]) - Access the ordered value in the window that is offset rows ahead of the current row. If there is no such row, then the default value will be returned. If not specified the offset is 1 and the default is null.

  • LAG(val [, offset [, default]]) - Access the ordered value in the window that is offset rows behind of the current row. If there is no such row, then the default value will be returned. If not specified the offset is 1 and the default is null.

  • NTH_VALUE(val, n) - Returns the nth val in window frame. The index must be greater than 0. If no such value exists, then null is returned.

Row value functions

  • ROW_NUMBER() – Sequentially assigns a number to each row in a partition starting at 1.

  • NTILE(n) – Divides the partition into n tiles that differ in size by at most 1. Larger tiles will be created sequentially starting at the first. n must be greater than 0.

Processing

Window functions are logically processed just before creating the output from the SELECT clause. Window functions can use nested aggregates if a GROUP BY clause is present. There is no guaranteed effect on the output ordering from the presence of window functions. The SELECT statement must have an ORDER BY clause to have a predictable ordering.

Note
 An ORDER BY in the OVER clause follows the same rules pushdown and processing rules as a top level ORDER BY. In general this means you should specify NULLS FIRST/LAST as null handling may differ between engine and pushdown processing. Also see the system properties controlling sort behavior if you different default behavior.

Teiid processes all window functions with the same window specification together. In general, a full pass over the row values coming into the SELECT clause is required for each unique window specification. For each window specification the values are grouped according to the PARTITION BY clause. If no PARTITION BY clause is specified, then the entire input is treated as a single partition.

The frame for the output value is determined based upon the definition of the analytical function or the ROWS/RANGE clause. The default frame is RANGE UNBOUNDED PRECEDING, which also implies the default end bound of CURRENT ROW. RANGE computes over a row and its peers together. ROWS computes over every row. Most analytical functions, such as ROW_NUMBER, have an implicit RANGE/ROWS - which is why a different one cannot be specified. For example, ROW_NUMBER() OVER (order)` can be expressed instead as count(*) OVER (order ROWS UNBOUNDED PRECEDING AND CURRENT ROW). Thus it assigns a different value to every row regardless of the number of peers.

Example: Windowed results
SELECT name, salary, max(salary) over (partition by name) as max_sal,
          rank() over (order by salary) as rank, dense_rank() over (order by salary) as dense_rank,
          row_number() over (order by salary) as row_num FROM employees
name salary max_sal rank dense_rank row_num

John

100000

100000

2

2

2

Henry

50000

50000

5

4

5

John

60000

100000

3

3

3

Suzie

60000

150000

3

3

4

Suzie

150000

150000

1

1

1

results matching ""

    No results matching ""