Re: real -> numeric -> real result different through jdbc

From: Michael Nacos <m(dot)nacos(at)gmail(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: pgsql-jdbc(at)postgresql(dot)org
Subject: Re: real -> numeric -> real result different through jdbc
Date: 2009-12-11 16:54:09
Message-ID: 407fa4640912110854y551f47d8q6ba4acb048cdd4f9@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-jdbc

2009/12/11 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>

> I seem to recall hearing that JDBC fools with the extra_float_digits
> setting. The "0.60000002" looks like it probably is the actual result
> of casting 0.6 to float4 --- remember float4 is only good to about six
> decimal digits. The default display of float4 rounds off at six digits,
> but in some contexts you don't want that rounding to happen because it
> might result in losing even more precision. JDBC is being conservative
> and taking the latter approach.
>

thanks tom,

we have just discovered this setting hidden away in
ConnectionFactoryImpl.java of the JDBC sources.
surely, it's not just a case of float4 display, type casting seems to be
affected, too.
does casting use the default display or does the internal representation
change?
It is causing our regression tests, which test for equality at the SQL
level, to fail.

cheers, Michael

In response to

Responses

Browse pgsql-jdbc by date

  From Date Subject
Next Message Michael Nacos 2009-12-11 17:22:19 Re: real -> numeric -> real result different through jdbc
Previous Message Tom Lane 2009-12-11 16:29:23 Re: real -> numeric -> real result different through jdbc