Creating list from list in python based on an if condition

Question

I want to create a new_list which will contain only the items of my old_list which satisfy the condition that the index in an array of labels is 3. I am trying something like this:

new_list = [x for x in old_list if idx[x] == 3]
IndexError: arrays used as indices must be of integer (or boolean) type

But i am getting the following error because idx is an array. How can i solve this problem?

edited: Idx is an array of equal size with my original data which contains labels for them. So basically i want to create a new list which will contain only the items of my original list which e.g have the label 3.

I want to do something like this: cluster_a = [old_list[x] for x in idx if x == 3]

Clarification: my old list is a list containing 3d arrays and the idx is an equal size array containing a label for each 3d array of my list as i aforementioned. I am trying my best to explain the problem. If something is needed please tell me.

This is the list with the 3d arrays

and this is the array with the labels

• Kasravnd: what you mean by index in an array of labels is 3 ? what index ? can you show an example ?
• azal: i mean the containing value
• kylieCatt: Can you show us what old_list looks like?
• Aida Paul: And why do you have to do it with list comprehension? By making it 3 lines longer this will become a very easy task, and at the same time code readability will skyrocket.
• chepner: Code readability will probably not be an issue if the correct code is written.

Answer 1

The problem is not that idx is a list but probably that x is an array - old_list must contain a list as an element. You need to reference the index, and not the item itself:

[old_list[x] for x in range(len(old_list)) if idx[x] == 3]

here's a minimal example:

>>> old_list = [4,5,6]
>>> idx = [3,2,3]
>>> [old_list[x] for x in range(len(old_list)) if idx[x] == 3]
[4, 6]

• chepner: This identifies the (likely) problem, but does not provide a solution.
• WeaselFox: The solution in this case is to rethink the logic. Its impossible to tell what can be done differently with the code exposed.
• chepner: Which is why you need clarification of the actual problem before providing an answer.
• Aida Paul: @WeaselFox which would mean that, currently, there is no possible answer and we should probably flag the question appropriately and/or ask in comments. But this isn't really an answer.
• WeaselFox: @Puciek - I dont agree, but in any case the question was updated, so that irrelevant.

Answer 2

What about this ? :

new_list = [x for x in old_list if idx.index(x) == 3 ]

• azal: when i try that i get an AttributeError:'numpy.ndarray' object has no attribute 'index'
• WeaselFox: idx.index(x) will return the first index in idx where the value is x, but according to the OP, idx is a list of labels, so I dont think this is right.