> > If I do the next : > > t1 = [(round(float(x),1), round(float(y),2)) for x, y in t] > > I get the long float as : > > [(35.799999999999997, -0.23999999999999999), (33.299999999999997, > -2.71), (33.600000000000001,-2.4199999999999999)] > > But I would have a float with 2 decimal numbers.
There is no such thing as a float with only two decimal numbers. This has been discussed on this ML a bazillion times - what you see above are rounding errors due to the approximation of decimal values by binary floating points. You can *convert a float to a string* and specify a precision when printing: >>> print "%.2f" % 2.4199999999999999999999999 2.42 Or you can use the module decimal to work with numbers base 10 - which you can also limit to certain precisions. See a recent thread on this ML, google for "Alternative to decimal type" Diez -- http://mail.python.org/mailman/listinfo/python-list
