Thread: CREATE AGGREGATE state function with one argument

In the manual fro creating aggregate functions (http://www.postgresql.org/idocs/index.php?sql-createaggregate.html) it
reads:

sfunc

     The name of the state transition function to be called for each input data value. This is normally a function of
twoarguments, the first being of type state_type and the second of type input_data_type. --->Alternatively, for an
aggregatethat does not examine its input values, the function takes just one argument of type state_type<---. In either
casethe function must return a value of type state_type. This function takes the current state value and the current
inputdata item, and returns the next state value. 

So I try this:

thalis=# CREATE FUNCTION state_func(int4) RETURNS int4 AS 'SELECT $1;' LANGUAGE 'sql';
CREATE
thalis=# CREATE AGGREGATE myaggr (basetype=int4,sfunc=state_func,stype=int4);
ERROR:  AggregateCreate: function 'state_func(int4, int4)' does not exist

I understand that I can have a state function that takes just one argument. Missing smthng?

TIA,
thalis

ps running 7.1.1

"Thalis A. Kalfigopoulos" <thalis@cs.pitt.edu> writes:
> In the manual fro creating aggregate functions (http://www.postgresql.org/idocs/index.php?sql-createaggregate.html)
itreads: 
> --->Alternatively, for an aggregate that does not examine its input
> values, the function takes just one argument of type
> state_type<---.

Yup.  It also says

input_data_type

     The input data type on which this aggregate function operates. This
     can be specified as ANY for an aggregate that
     does not examine its input values (an example is count(*)).

Perhaps "can" is the wrong word there, you *must* say ANY if your
aggregate is input-value-independent.

            regards, tom lane