numpy and python array index question

[jenya56]

Dear all:

I have this function and it works just fine but it is slow:

def common_index(list1, list2, val1, val2):
   assert len(list1) == len(list2), "List not of equal length"
   for i in range(len(list1)):
      if list1[i] == val1 and list2[i] == val2:
         return [i] 
To optimize it:
coord_ind=numpy.where(list1==val1 and list2==val2) 

Yet this line does not work and always returns ()? Any suggestions? Thanks!!!

[Glenton]

As with my other post you probably want to be working in numpy. And using what I believe are called masks is a great idea.

We'll change list1 and list2 to array1 and array2. We're looking for a list of the indexes where array1[i]==val1 and array2[i]==val2.

indices=arange(n)   #this creates an array of the indices you're after
mask1=array1==val1  #this creates an array like [True, False, False, True,...]
mask2=array2[mask1]==val2  #(no need to check all of the value!)
                                   #array2[mask1] is a smaller array
return indices[mask1][mask2]
 

I haven't tried it yet, but it *should* work, and it *should* be a ton quicker...
[/code]