Hi Larry,
If I understand correctly, then the use of CTE in row based database would be beneficial as it would form a view on the selected columns rather than the entire row data. The materialization of the view may happen in memory. This would help speed up the performance.
But in case of HANA, anyway the data read is column based. The materialization of the view would anyway happen for the select statement. so I don't think it would add any additional benefits.
Secondly as it has been correctly mentioned, the CE functions would provide much better performance.
I'd rather wish more analytic functions to be integrated, exception handling to be declared and provide more robust handling of error situations.
Regards,
Ravi