Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!i2unix!irst!ferrario
From: ferrario@irst.UUCP (Lorenza Ferrario)
Newsgroups: comp.databases
Subject: INGRES SELECT HELP
Message-ID: <329@irst.UUCP>
Date: 4 Dec 89 09:49:00 GMT
Organization: I. R. S. T. , Trento - Italy
Lines: 56

I'm trying to understand if it is possible to build an INGRES ESQL-C procedure
to allow to create a "select" with variable columns, that is, passing the name
of columns as program's variables (:name);
they follow pieces of the code I wrote to solve this problem.

...
exec sql begin declare section;
char *vect[20], *trans, tmpcol[20];
...
exec sql end declare section;

...

vect[0]="col1"; vect[1]="col2"; ....

...

for(i=0; i<i;){
	trans=vect[i];
	exec sql declare tmpcur cursor for
	select :trans
	from tabella;
	
exec sql open tmpcur;
exec sql whenever not found goto tmpcur_done;
while(sqlca.sqlcode == 0){
	exec sql fetch tmpcur into
	:tmpcol;
	
	exec frs loadtable scelta tabellaf  /* "scelta", "tabellaf" are form's
                                               name and tablefield's name 
					       respectively */
	(:trans = :tmpcol);
	}
	tmpcur_done:
	exec sql close tmp_cur;
++i;
}

...  	 	

It doesn't work as I aspected because the "select" section handles the column
name (:trans) as a constant string , not as a vairiable,so in the "tabellaf" 
tablefield's column they appear "col1", "col2" and not the stored values of 
these columns in the db table "tabella";  meanwhile in the "loadtable"
section there are no problems.

Is it possible to solve all that or do you think it is a stupid work?

		Thanks in advance

				lorenza ferrario

-- 

	Lorenza Ferrario