Thread: Float8 precision problem

And (indeed) that is exactly the answer that you received [within
DLB_DIG units of precision].

I guess that you will be happier with NUMERIC(precision, scale) because
the results of operations will be closer to what you expect.

Suggested reading:
http://www.physics.ohio-state.edu/~dws/grouplinks/floating_point_math.pd
f

________________________________________
From: pgsql-general-owner@postgresql.org
[mailto:pgsql-general-owner@postgresql.org] On Behalf Of Hengki
Suhartoyo
Sent: Thursday, November 30, 2006 2:56 AM
To: pgsql-general@postgresql.org
Subject: [GENERAL] Float8 precision problem

Hi all,

I got problem in postgrsql 8.0 when subtraction in float8, this is my
query :
select 6.1::float8-6::float8

Result>>> 0.099999999999996

I need the result like natural subtraction is 0.1

Help me please....

________________________________________
Want to start your own business? Learn how on Yahoo! Small Business.

On Nov 30, 2006, at 19:56 , Hengki Suhartoyo wrote:

> I got problem in postgrsql 8.0 when subtraction in float8, this is
> my query :
> select 6.1::float8-6::float8
>
> Result>>> 0.099999999999996
>
> I need the result like natural subtraction is 0.1

Then use numeric instead of float. Float by its very nature is not
precise.

Michael Glaesemann
grzm seespotcode net

Hi,

What about cast to numeric?

select 6.1::numeric-6::numeric ;
 ?column?
----------
      0.1
(1 row)


Regards,
  Kaloyan Iliev

Hengki Suhartoyo wrote:

> Hi all,
>
> I got problem in postgrsql 8.0 when subtraction in float8, this is my
> query :
> select 6.1::float8-6::float8
>
> Result>>> 0.099999999999996
>
> I need the result like natural subtraction is 0.1
>
> Help me please....
>
> ------------------------------------------------------------------------
> Want to start your own business? Learn how on Yahoo! Small Business.
> <http://us.rd.yahoo.com/evt=41244/*http://smallbusiness.yahoo.com/r-index>