Archives of the TeradataForum

Message Posted: Wed, 31 Oct 2001 @ 15:47:38 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

In the Data Warehouse we are currently working on, we have defined all the columns to be NOT NULL.

In our mind, some of the advantages of definining the columns as "NOT NULL" were:

1. The users would not have to worry about whether the column can have NULL values or not. Hence, they won't have to worry about including the "IS NULL" keywords in the WHERE clause. (Ease of use)

2. The use of the IS NULL keywords in the WHERE clause makes that portion of the query non-sortable, which means that portion of the query cannot use an index. (Increased performance)

Now, the customer is asking us to change most of the non-primary key NOT NULL columns to NULL columns to make is consistent with another Data Mart that they have. Please note that some of the fields will never have NULL values, but the customer is insisting that we still define them as NULL.

Are there any other performance/storage issues if we define fields as NULL instead of NOT NULL?

I would highly appreciate any input on these issues so that we can convince the customer that changing the non-primary key columns to NULL is not a good idea.

Thanks in advance for your help.

Best Regards,

Bharat

	<Prev	Next>		<<First	<Prev	Next>	Last>>