[SciPy-user] amax, amin, mean

[Lionel Roubeyrie]

Sorry but I don't understand how to do, replacing nan by an number gives 
others errors:

######################################################
lionel[~]17>tv=[1,2,3,4,5,1.e-20,6]

lionel[~]18>tv=MA.masked_values(tv, 1.e-20)

lionel[~]19>amax(tv)
---------------------------------------------------------------------------
MA.MA.MAError                                   Traceback (most recent call 
last)

/home/lionel/<console>

/usr/lib/python2.4/site-packages/scipy_base/function_base.py in amax(m, axis)
    179         axis = 0
    180     else:
--> 181         m = _asarray1d(m)
    182     return maximum.reduce(m,axis)
    183

/usr/lib/python2.4/site-packages/scipy_base/function_base.py in 
_asarray1d(arr)
    152     """Ensure 1d array for one array.
    153     """
--> 154     m = asarray(arr)
    155     if len(m.shape)==0:
    156         m = reshape(m,(1,))

/usr/lib/python2.4/site-packages/scipy_base/type_check.py in asarray(a, 
typecode, savespace)
     23             r.savespace(savespace)
     24          return r
---> 25    return multiarray.array(a,typecode,copy=0,savespace=savespace or 0)
     26
     27 ScalarType = [types.IntType, types.LongType, types.FloatType, 
types.ComplexType]

/usr/lib/python2.4/site-packages/Numeric/MA/MA.py in __array__(self, t)
    630         if self._mask is not None:
    631             if Numeric.sometrue(Numeric.ravel(self._mask)):
--> 632                 raise MAError, \
    633                 """Cannot automatically convert masked array to 
Numeric because data
    634                    is masked in one or more locations.

MAError: Cannot automatically convert masked array to Numeric because data
                   is masked in one or more locations.

######################################################

What do I have to do for having real computation on masked arrays?
Thanks

-L. Roubeyrie