The SINGLE predicate tests whether or not the result table specified in the column select expression has exactly one row. If it has exactly one row, SQL evaluates the SINGLE predicate as true. If the result table has zero rows or more than one row, the predicate is false. SQL evaluates the NOT SINGLE predicate as true if the result table specified in the select expression has zero rows or more than one row. The SINGLE and NOT SINGLE predicates cannot be unknown. The SINGLE predicate has the following form: (B)0[m[1msingle-predicate = [m [1m [m [1mqqq> [1;4mSINGLE[m[1m qq> ( select-expr ) qqqqq>[m Because it checks only for the existence of rows, a SINGLE predicate does not require that the result table from its column select expression be a single column wide (see Column_Select_ Expressions for details on column select expressions). For SINGLE predicates, an asterisk (*) wildcard in the column select expression can refer to a multicolumn table (as in the following example).