My question would be: why an index range scan MUST be used? What is the reason
for this person to deem it mandatory?
This would also discard unique index access, which is a different access path
than an index range scan.
I question the knowledge of the person if he/she can not see the point of a
full table scan or any other different access path.
Even if the documentation would state that an index must be used, and does not
provide a valid reason for why it shouldn’t be used, I would discard the advise
in the documentation. There are lots of examples where the official
documentation is incorrect, and even experts might be wrong; for something to
be written in a book doesn’t make it the truth.
Mobile: +31 6 14180860
On 10 Apr 2018, at 18:13, Ls Cheng <exriscer@xxxxxxxxx> wrote:
I have a Dinosaur stating we must always use index range scan no matter OLTP
or DWH systems. This guys still lives in the late 90's and Oracle 7 (although
he claims he is 18 years veteran expert from 7 up to 12c), the thing is I
cannot find any Oracle WP or similar in the docs which states Full Table Scan
is not evil except Tom Kyte (asktom). I have even demo-ed him with a very
simple two table join example with FTS and index range effect when a large
amount of data need to be accessed, still he says in Oracle docs states that
scattered reads (FTS) mist be reduced!
Does anyone know any official pointer, in WP form or MOS support note which
talks about FTS vs index scans?