Why a data structure subfield with the same name of a file field does not keep the same data type?

Hi all,
I am struggling with a subfield definition issue which I haven't understood
yet, and it's driving me crazy!

In a program, an update keyed physical file is defined with a packed (5, 0)
key field.
The file definition is global, like this:
dcl-f tblstr disk usage(*input) keyed;
The program passes that key field plus a second field to a procedure, and
the procedure uses those fields in an embedded SQL, like this:
dcl-proc UpdFile;
dcl-pi *n zoned(5:0);
x_STRNUM like(STRNUM) value;
x_STRNAM like(STRNAM) value;
end-pi;
dcl-s count zoned(5:0) inz(0);
message = '4-' + %char(x_STRNUM) + %trim(x_STRNAM);
dsply message ' ';
exec sql
select count(*)
into :count
from tblstr
where STRNUM = :x_STRNUM and STRNAM = :x_STRNAM
;
return count;
end-proc;


Well, the program doesn't compile, because both x_STRNUM and x_STRNAM are
not usable by the SQL statement:
Position 21 Variable X_STRNUM not defined or not usable.
Reason: No declaration for the variable exists, the
declaration is not within the current scope, or the
variable does not have an equivalent SQL data type.

The same for x_STRNAM.

When the SQL 'where' clause is commented, the program compiles and it
works, meaning the 'dsply' operation code is correctly executed.

I tried different options, like adding value to the parameters passed and
casting the rpg variables used within the SQL statement, but I can't get
rid of the issue.

The idea I get from all the tries is that file's fields used to define the
input parameters of the procedure are not recognized by SQL, so the
variables inside the procedure are for some reason unusable.

What am I missing? Does any of you have any idea about it?

Any hint or suggestion is greatly appreciated as converting fixed RPG
programs to free format includes many of these situation.

Thanks!
Lucia