Group by problem

Question

I need to group by ,even if is there a difference in values for a column by converting to a text value and needs that text value to be included in select list

How is this possibe?

  SELECT col1,col2
   FROM
       (
         SELECT col1,col2 FROM table1
         UNION ALL
     SELECT col1,col2 FROM table2
         UNION ALL
         SELECT col1,col2 FROM table3

       )tbl
  GROUP BY  tbl.col1
           ,tbl.col2

here col2 may or may not be a text value or int, if it is text value i need to convert all other col2 ints to text and do a group by

See this example :

Table 1
-------
A B  3   C 
A B  var C


Table 2
-------
UNION ALL

B B  3   C 
B B  var C


Table 3
-------
UNION ALL

B B  3   C 
B B  3   C

Result of each table should be 
-------

Table 1
-------
A B  var C (since there is a var in the any of the row in that column in that table)

Table 2
-------
B B  var C (since there is a var in the any of the row in that column in that table)

Table 3
-------
B B  3   C  (here it remains 3 since there is not value call var in any of the row in that column)

Result 
------
A B  var C
B B  var C  (since there is a var in the any of the row in that column in that table)

Answer 1

This seems a little to easy, maybe i am missing something? Are you just looking to do a CAST on col2?

SELECT  col1
       ,CAST(col2 AS nvarchar(100)) as Col2
       ...etc

Answer 2

tAKE THIS:

 SELECT col1,col2
   FROM
       (
         SELECT col1,col2 FROM table1
         UNION ALL
     SELECT col1,col2 FROM table2
         UNION ALL
         SELECT col1,col2 FROM table3

       )tbl
where col2 not like '3%'
  GROUP BY  tbl.col1

The Result look like your question:

A B var C

B B var C