It can be very dangerous to call functions you did not write yourself. you have no control over the datatypes!

You want to use BINARY_INTEGER, not the Oracle Number type (very doubtful windows would be using an OCINumber :)

Also, the function is really GetPrivateProfileStringA <<=== A? why A? who knows....

Here is a full example:


tkyte@ORA8I> create or replace library kernel32 as 'C:\windows\system32\kernel32
.dll';
2 /

Library created.

tkyte@ORA8I>
tkyte@ORA8I> set echo on
tkyte@ORA8I>
tkyte@ORA8I>
tkyte@ORA8I> create or replace
2 function pGetPrivateProfileString
3 (p_section in varchar2,
4 p_key in varchar2,
5 p_default in varchar2,
6 p_returnvalue out varchar2,
7 p_size in BINARY_INTEGER,
8 p_FileName in varchar2)
9 return BINARY_INTEGER
10 as language c
11 library kernel32
12 name "GetPrivateProfileStringA";
13 /

Function created.

tkyte@ORA8I>
tkyte@ORA8I> declare
2 str_out varchar2(255);
3 n number;
4 begin
5 n := pGetPrivateProfileString
6 ('SAMPLE_LIST',
7 'NameOfSample',
8 'NameOfSampleDefault',
9 str_out,255,
10 'C:\TEMP\test.ini');
11 dbms_output.put_line( 'n = ' || n );
12 dbms_output.put_line( 'str_out = ' || str_out );
13 end;
14 /
n = 19
str_out = NameOfSampleDefault

PL/SQL procedure successfully completed.

tkyte@ORA8I>
tkyte@ORA8I> host echo [sample_list] > c:\temp\test.ini

tkyte@ORA8I> host echo NameOfSample = foobar >> c:\temp\test.ini

tkyte@ORA8I> /
n = 6
str_out = foobar

PL/SQL procedure successfully completed.