This post is to give an answer to Simon Slavin on the mailing list (unfortunately the buton "reply via email to Simon Slavin" on the web access to the mailing list doen't work for me (firefox 3.6.16) and returns "Requested archive does not exist" error)
The method of rounding floating point values we need is the (most ?) common one and could be described as : "increment the last digit in the range given by the precision argument if the following digit (if existing) is >= 5". For instance round(8.875,2) will return 8.88, round(16.875,2) will return 16.88 etc... In our country (France), this method is sometime called "commercial rounding". This method is much simpler and less "sophisticated" than the "banker's rounding" method. In our country the "commercial rounding" method is one of the two accepted methods by Tax Services (the other one, depending on the circomstances, beeing the truncation of the floating point value and retaining only the integer part). I believe the intended method followed by the round(x,y) function in SQLite is the "commercial rounding" method. Jean-Marie _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
