I have the following arrays:
a = [10, 31, 30, 11, 17, 12, 22, 25, 85, 17, 21, 43]
b = [0, 1, 4, 6]
I want to mask a based on the index given by array b. That means getting:
c = [True, True, False, False, True, False, True, False, False, False, False, False]
In [524]: a=np.array([10, 31, 30, 11, 17, 12, 22, 25, 85, 17, 21, 43])
In [525]: b=np.array([0, 1, 4, 6])
To make a boolean c that is True at the b indices, just use:
In [526]: c=np.zeros(a.shape, bool)
In [527]: c[b]=True
In [528]: c
Out[528]:
array([ True, True, False, False, True, False, True, False, False,
False, False, False], dtype=bool)
Then you can select the values of a with:
In [529]: a[c]
Out[529]: array([10, 31, 17, 22])
but you could just as well select them with b:
In [530]: a[b]
Out[530]: array([10, 31, 17, 22])
but c is better for removing those, a[~c]. np.delete(a,b) does the same thing.
Other array methods of generating c are
np.in1d(np.arange(a.shape[0]),b)
np.any(np.arange(a.shape[0])==b[:,None],0)
And since I was just discussing masked arrays in another question, I could do the same here:
In [542]: np.ma.MaskedArray(a,c)
Out[542]:
masked_array(data = [-- -- 30 11 -- 12 -- 25 85 17 21 43],
mask = [ True True False False True False True False False False False False],
fill_value = 999999)