Le 2013-03-06 à 21:42, Tony Dare a écrit : > I'm taking an standard deviation of a population and subtracting it from the average of the same population and rounding the result. Sometimes that result is negative and rounding it returns (or shows up as) a negative zero (-0) in a SELECT. > > basically: > SELECT > client_name, avg(rpt_cnt), > stddev_pop(rpt_cnt), > round(avg(rpt_cnt) - stddev_pop(rpt_cnt)) > from client_counts > group by client_name > > and what I sometimes get is : > client_name | a dp number | a dp number | -0 > > In postgresql-world, is -0 = 0? Can I use that negative 0 in further calculations without fear? Is this a bug? This is related to the recent discussion of floating point values on this mailing list. You can read more about IEEE 754 and whether 0 == -0 on Wikipedia: https://en.wikipedia.org/wiki/Signed_zero#Comparisons According to that article, IEEE 754 specifies that 0 == -0 in Java/C/etc. Hope that helps! François Beausoleil
<<attachment: smime.p7s>>
