Python: numpy array sublist match with large list where sequence matter

One very easy solution using the sublist pattern for slicing and checking for that signature and getting the indices would be -

np.flatnonzero(~a[:-2] & a[1:-1] & a[2:]) # a as data array

Explanation

Basically, we are slicing three slices out of the data array - One that starts at 0th index and goes until leaving out last two elements, another that starts at 1st index and ends at second last element and a third slice starting at 2nd and goes on until the end. These three slices correspond to the three elements of matching required against the sublist pattern that's - [False, True, True]. We need to make sure that the first one is False, in other words, let's make sure the negation of it is True. Negation in NumPy is achieved through ~ operator. So, in essence, we get the combined mask off those three slices and get the corresponding indices with np.flatnonzero.

For the given data results in -

In [79]: np.flatnonzero(~a[:-2] & a[1:-1] & a[2:])
Out[79]: array([ 12,  31,  68, 112])

in won't check for sub-arrays. Instead, it checks for elements.

You will have to do something like this:
(Using A for big array and b for sub-list for readability.)

n = len(b)
c = [i for i in xrange(len(A)-n+1) if (b==A[i:i+n]).all()]

c is the required list of indexes.

Explanation:
This is basic List Comprehension in python.
The idea is to create sub-arrays of the bigarray and check if it matches the sublist.

Breaking down the statement for better understanding:

c = []    
for i in xrange(len(A)-n+1):
    if (b==A[i:i+n]).all():    # if list and arrays match
        c.append(i)