Oracle SQL Operator >= has higher performance than >

Question

I got asked one question in interview.

Below two SQL will get same result. but which one have higher performance?

SELECT * FROM EMP WHERE DEPTNO >= 4

SELECT * FROM EMP WHERE DEPTNO > 3

The answer is first one . In first SQL , Database will locate DEPT =4 directly where searching. it has higher performance. In second SQL , Database will locate DEPT =3, and forwardly scan row which is larger than 3 .

It is my first time heard that theory. I can't see any difference form execute plan and stat number.

Is there any official explanation for it ?

I found some Chinese website share the same tips.

http://edm.ares.com.tw/dm/newsletter-2014-03-uPKI-OTP-newrelease/it-1.php

Answer 1

I find that to be a myth if the column has an index on that field. I tried with a primary key and both execution plans use an Index Range Scan which is quite fast on a primary key or unique index. The two queries performance is extremely close to each other and both return a similar cost, but the one with ">" was 0.3% better on the estimated CPU cost than ">=". I say BUSTED!

Besides, when I say something performs better, normally I would mean 10%+ if not more.