Ranking functions return a value for each row in the input. The ranking functions supported by SQL Anywhere are RANK, DENSE_RANK, PERCENT_RANK, and CUME_DIST. Ranking functions are not considered aggregate functions because they do not compute a result from multiple input rows in the same manner as, for example, the SUM aggregate function. Rather, each of these functions computes the rank, or relative ordering, of a row within a partition based on the value of a particular expression. Each set of rows within a partition is ranked independently; if the OVER clause does not contain a PARTITION BY clause, the entire input is treated as a single partition. Consequently, you cannot specify a ROWS or RANGE clause for a window used by a ranking function. It is possible to form a query containing multiple ranking functions, each of which partition or sort the input rows differently.
All ranking functions require an ORDER BY clause to specify the sort order of the input rows upon which the ranking functions depend. If the ORDER BY clause includes multiple expressions, the second and subsequent expressions are used to break ties if the first expression has the same value in adjacent rows. NULL values in SQL Anywhere are always sorted before any other value (in ascending sequence).
|Send feedback about this page via email or DocCommentXchange||Copyright © 2008, iAnywhere Solutions, Inc. - SQL Anywhere 11.0.0|