Re: 9.5 "chained equality" behavior

> On 30 May 2017, at 22:56, Joshua Ma <josh(at)benchling(dot)com> wrote:
>
> Our team is upgrading from 9.4 to 9.5, and we noticed this behavior change:
>
> 9.4:
> # SELECT true = true = true;
> ?column?
> ----------
> t
> (1 row)

Does that really do what you intended though?

# select false = false = false;
?column?
----------
f
(1 row)

I suspect you expect that to result in 't', not 'f', so replacing that with AND would change the behaviour for false boolean values. (This is from 9.3, btw)

Perhaps explicitly adding the braces that pre-9.5 pg adds implicitly gets you what you had? I suspect that would work on 9.5 as well (but can't test right now).

# select false = (false = false);
?column?
----------
f
(1 row)

# select true = (true = true);
?column?
----------
t
(1 row)

Alban Hertroys
--
If you can't see the forest for the trees,
cut the trees and you'll find there is no forest.