You can reference all variables declared in an Embedded SQL declaration section and the preprocessor accepts them from the point of declaration to the end of the file. This may not be true for the Fortran compiler, which allows references to variables only in the scope of the program unit in which they were declared. If you have two unrelated subprograms in the same file, each of which contains a variable with the same name to be used by Embedded SQL, do not redeclare the variable to Embedded SQL. The preprocessor uses the data type information supplied the first time you declared the variable.

In the following program fragment, the variable dbname is passed as a parameter between two subroutines. In the first subroutine, the variable is a local variable. In the second subroutine, the variable is a formal parameter passed as a string to be used with the connect statement. The declaration of dbname in the second subroutine must not occur in an Embedded SQL declaration section. In both subroutines, the preprocessor uses the data attributes from the variable's declaration in the first subroutine.

Take special care when using variables in a DECLARE CURSOR statement. The variables used in such a statement must also be valid in the scope of the OPEN statement for that same cursor. The preprocessor actually generates the code for the declare at the point that the OPEN is issued and, at that time, evaluates any associated variables.

For example, in the following program fragment, even though the variable "number" is valid to the preprocessor at the point of both the DECLARE CURSOR and OPEN statements, it is not an explicitly declared variable name for the Fortran compiler at the point that the OPEN is issued, possibly resulting in a runtime error. Because Fortran allows implicit variable declarations (although Embedded SQL does not), the compiler itself does not generate an error message.

You must issue the INCLUDE SQLCA statement in each subprogram that contains Embedded SQL statements.