python lambda list filtering with multiple conditions

Question

My understanding about filtering lists with lambda is that the filter will return all the elements of the list that return True for the lambda function. In that case, for the following code,

inputlist = []
inputlist.append(["1", "2", "3", "a"])
inputlist.append(["4", "5", "6", "b"])
inputlist.append(["1", "2", "4", "c"])
inputlist.append(["4", "5", "7", "d"])

outputlist = filter(lambda x: (x[0] != "1" and x[1] != "2" and x[2] != "3"), inputlist)
for item in outputlist: print(item)

The output should be

['4', '5', '6', 'b']
['1', '2', '4', 'c']
['4', '5', '7', 'd']

But the output that I get is

['4', '5', '6', 'b']
['4', '5', '7', 'd']

I get the expected output, if I use

outputlist = filter(lambda x: (x[0] != "1" or x[1] != "2" or x[2] != "3"), inputlist)

What am I doing silly here? Or is my understanding not correct?

Answer 1

x = ['1', '2', '4', 'c'], so x[1]=='2', which makes the expression (x[0] != "1" and x[1] != "2" and x[2] != "3") be evaluated as False.

When conditions are joined by and, they return True only if all conditions are True, and if they are joined by or, they return True when the first among them is evaluated to be True.

Answer 2

['1', '2', '4', 'c']

Fails for condition

x[0] != "1"

as well as

x[1] != "2"

Instead of using or, I believe the more natural and readable way is:

lambda x: (x[0], x[1], x[2]) != ('1','2','3')

Out of curiosity, I compared three methods of, er... comparing, and the results were as expected: slicing lists was the slowest, using tuples was faster, and using boolean operators was the fastest. More precisely, the three approaches compared were

list_slice_compare = lambda x: x[:3] != [1,2,3]

tuple_compare = lambda x: (x[0],x[1],x[2]) != (1,2,3)

bool_op_compare = lambda x: x[0]!= 1 or x[1] != 2 or x[2]!= 3

And the results, respectively:

In [30]: timeit.Timer(setup="import timeit,random; rand_list = [random.randint(1,9) for _ in range(4)]; list_slice_compare = lambda x: x[:3] != [1,2,3]", stmt="list_slice_compare(rand_list)").repeat()
Out[30]: [0.3207617177499742, 0.3230015148823213, 0.31987868894918847]

In [31]: timeit.Timer(setup="import timeit,random; rand_list = [random.randint(1,9) for _ in range(4)]; tuple_compare = lambda x: (x[0],x[1],x[2]) != (1,2,3)", stmt="tuple_compare(rand_list)").repeat()
Out[31]: [0.2399928924012329, 0.23692036176475995, 0.2369164465619633]

In [32]: timeit.Timer(setup="import timeit,random; rand_list = [random.randint(1,9) for _ in range(4)]; bool_op_compare = lambda x: x[0]!= 1 or x[1] != 2 or x[2]!= 3", stmt="bool_op_compare(rand_list)").repeat()
Out[32]: [0.144389363900018, 0.1452672728203197, 0.1431527621755322]