Archives of the TeradataForum

Message Posted: Mon, 21 Jul 2003 @ 14:03:41 GMT

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

The DBC tables that use INTEGER to store times are a historical anomaly. Those tables were "invented" before Teradata had a TIME datatype, and so time values were typically stored either in INTEGER or FLOAT columns. There was the built-in system function TIME (now better called CURRENT_TIME) that returned the time as a FLOAT.

If you didn't need the fractional seconds, storing it in an INTEGER cut the space use in half.

If there had been a TIME datatype back then, we would certainly have used it in those DBC tables.

I assume the reason that TIME(0) isn't optimized for space use is that:

1) for most users, it's not an issue.

2) It greatly simplifies the processing internal to the system if the TIME is always kept in the same format.

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