Re: 'real' strange problem in 7.1.3

reina(at)nsi(dot)edu (Tony Reina) writes:

> db02=# select distinct arm from ellipse where exag_ratio = 1.0;

You never want to use the = test on floating point numbers. Two
apparently equal numbers may differ in the least significant digit.
The behavior will be close to random.

When you convert the floating point to a string and round off to a
specific number of digits, you can use string compare to get more
predictable results.

Another possibility is to do something like

where abs(exag_ratio - 1.0) > 0.000001

(I'm not sure about the SQL function for absolute value, but you get
the idea).

-Knut

--
The early worm gets the bird.