>Ok, here's the problem with this.
>
>I'm working on a legacy app. I'm writing a process that automatically inserts records into
>a DBF.
>
>There is a default value for the PK field:
>
>

new_pk("custpoc","poc_pk")

>
>which calls a stored proc that returns the row's PK value.
>
>My process forms an INSERT statement, which is throwing an error "Uniqueness of index POC_PK is violated."
>
>Not all tables are calling this stored proc, so this presents a problem.
>
>I'm trying to make my process generic so that I can use it in any app. I need a way to
>exclude the PK column from the insert statement when this rule is in place.

I'd first look at the existing legacy source to see how INSERTs are handled elsewhere. If the pkey column has a default value then it's not an autoincrement field so you can't filter it out that way.

Can you just use the function call as the value you're INSERTing? i.e.

INSERT INTO CustPOC ;
  ( POC_PK ;
  , ... ) ;
  VALUES ;
    ( New_PK( "CustPOC", "POC_PK" ) ;
    , ... )