Re: Underlying Implementation of VARCHAR data types

mikelbell2000_at_yahoo.com (Mike L. Bell) wrote in message news:<49d7474a.0403080855.70089aa7_at_posting.google.com>...
> ...
> Also, in Oracle do I not have to worry about performance penalties
> associated with VARCHAR types in other systems with things like page
> overflows if the resultant update of the field results in a length
> that will no longer fit on the same data page?

We often see debate on CHAR vs VARCHAR2 (when the column is known to contain fixed length strings). I don't recall seeing anybody's performance test. If there is one, the difference may be barely noticeable.

But I remember a few years ago I looked at the location of data inside a data block and found that updating 'NY' to 'TX' or even to 'NY' could change the data location in the block if the column is VARCHAR2. If it's CHAR, it stays in the same location. I could remember wrong but if that's the case, it may have performance implication.

Yong Huang Received on Mon Mar 08 2004 - 22:22:45 CST