sql - How to make GROUP BY as a parameter without using CASE WHEN? -

I have the following table with the following tables and values ​​and types.

  Create a table example (fname text, lname text, value int); Enter the values ​​in the example ('doe', 'coin', 123), ('bit', 'coin', 434), ('light', 'coin', 565), ('dogs', 'mem' 183), ('bit', 'mem', 453), ('light', 'mem', 433); Create (name text, zodiac int) as a typed result;  

I would like to write a function, which is given by a function to the group. This example works:

  declare text $$ my_parameter; Result result []; Start my_parameter = 'last'; Result: = array (when my_premator = 'first' then fender occurs when my_parameter = 'last' then lname end, sum (pay)): Example group from example group when my_parameter = 'first' then fname when my_parameter = 'last 'And then' lname 'end); Increase notices '%', results; End; $$ language plpgsql;  

I have been told that when a decision is really expensive, a clear solution is to make the selectors twice:

  if my_parameter = 'First' then result: = array (selection line (fname, sum (pay)): resulting from example group fname; end if; If my_parameter = 'last' then the result: = array (selection line (lname, sum (salary)): result from example group by lname; if end;  

but it does not change the ugly duplicate code

Is there another way to make the group eligible for the parameter?

If you do not want to use the case , you can use it:

  Choose whistle (name, pay) (e.g. , Where my_parameter = 'first' is the choice of a union Do not, where my_parameter = 'last') salary name, / code>  
 But, in fact, it is better to check, I have not heard that the  case  is expensive . 
 
 If you think that  case  is not expensive, then I still suggest subquery or cte to avoid code duplicates, like: 
 < Pre-> with  as the CT (name, pay) name (select my_parameter = 'first' then fname when my_parameter = 'last' then name as name, salary, for example) name not Gm, zodiac sign (salary) from whistle group