Archives of the TeradataForum

Message Posted: Thu, 25 Sep 2003 @ 22:14:51 GMT

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

We have some history tables in our datamart that have columns eff_dt and exp_dt. We have some monthly metrics table as well, which has a date column (supposed to hold the last day of the month).

Most of the times the metrics table(mt) is joined with these history tables. So say for example, I need a particular attribute in the history table (ht) as of a date in the metrics table, my sql will look something like this.

Select ...
Where ht.Eff_dt <=mt.as_of_date and ht.exp_dt > mt.as_of_date.

Now it is being suggested that all these date cols should be of numeric datatype. (I do not agree.) The reasoning being that numeric datatypes perform well in joins compared to date types.

Now for the questions.

1. Is this true.

2. Is there any documentation to prove this.

Would much appreciate any response.

Thanks!!!

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