Thread: Exponential notation bug

I took a look at simplifying this test case, and I think it comes down
to data types:

maciek=# select 9/10;
 ?column?
----------
        0
(1 row)

maciek=# select pg_typeof(10);
 pg_typeof
-----------
 integer
(1 row)

But:

maciek=# select 9/1e1;
        ?column?
------------------------
 0.90000000000000000000
(1 row)

maciek=# select pg_typeof(1e1);
 pg_typeof
-----------
 numeric
(1 row)

Does that explain the behavior you're seeing?

On Fri, Apr 4, 2025 at 11:55 AM Marcos Pegoraro <marcos@f10.com.br> wrote:
> I was deleting thousands of records each time with \gexec, then ...
>
> This one works
> select format('delete from table where ID = any(%L::integer[]);', array_agg(ID)), (ord-1)/10000 from (
> select * from generate_series(15e2,65e5) with ordinality) x(ID, ord) group by 2 order by 2;
>
> But it's easier to write in exponential notation, so I changed this (ord-1)/10000 to (ord-1)/1e4. Using this way
exponentialnotation is just ignored. 

This seems like a question for -general or some other user-focused
mailing list, not hackers. At any rate, I don't see how 1e4 could just
be "ignored", but as Maciek points out, 1e4 and 10000 are of different
data types, which seems likely to be relevant somehow.

--
Robert Haas
EDB: http://www.enterprisedb.com

Robert Haas <robertmhaas@gmail.com> writes:
> This seems like a question for -general or some other user-focused
> mailing list, not hackers. At any rate, I don't see how 1e4 could just
> be "ignored", but as Maciek points out, 1e4 and 10000 are of different
> data types, which seems likely to be relevant somehow.

I am also wondering about careless whitespace.

postgres=# select 9/1e1;
        ?column?        
------------------------
 0.90000000000000000000
(1 row)

postgres=# select 9/1 e1;
 e1 
----
  9
(1 row)

In any case, I tried copying-and-pasting parts of the originally
mentioned query, and I saw nothing that looked like misbehavior.
Please show an exact, self-contained test case if you want us
to believe there's a bug here.

            regards, tom lane