Teradata: On doing PIVOT operation values of columns interchange - sql

Question

I am having a Table (DB.TAB_UNPIVOTED) on TeraData having millions of rows, and 4 columns - ID, Week, Sales & Profits. Here below is a small microcosm of the problem.
For sake of illustration DB.TAB_UNPIVOTED is like this:
I want to PIVOT this table. Following is my code:
SET SQL_STMT = ('CREATE TABLE DB.TAB_PIVOTED AS
( SELECT * FROM DB.TAB_UNPIVOTED
PIVOT
( SUM(Sales) AS Sales, SUM(Profits) AS Profits
FOR KW_Prefix IN (
'CW_1' AS CW_1,
'CW_2' AS CW_2,
'CW_3' AS CW_3
)
) AS dt
) WITH DATA;'
) ;
EXECUTE IMMEDIATE SQL_STMT;
The strange thing is that, on PIVOTING, everytime I get a different table DB.TAB_PIVOTED, where values across different columns are interchanged. For eg; one time the output can be:
But, next time, it could be a dfferent result, with values interchanged, though Profits & Sales maintain their pairing. In one output they could be under CW_1, where as on another run, it could be under CW_3:
This problem may not be reproduced on small data, but with my data in millions and CW varing from 1 to 52, I see it all the time.
Does anyone have an idea, where in the pivoting am I making the mistake?
Inputs would be very esteemed.
Update:
I tried running the code directly as well, instead of EXECUTE IMMEDIATE SQL_STMT,but same strange results.
CREATE TABLE DB.TAB_PIVOTED AS
( SELECT * FROM DB.TAB_UNPIVOTED
PIVOT
( SUM(Sales) AS Sales, SUM(Profits) AS Profits
FOR KW_Prefix IN (
'CW_1' AS CW_1,
'CW_2' AS CW_2,
'CW_3' AS CW_3
)
) AS dt
) WITH DATA;