5

I have a list of tuples that is created with the zip function. zip is bringing together four lists: narrative, subject, activity, and filer, each of which is just a list of 0s and 1s. Let's say those four lists look like this:

narrative = [0, 0, 0, 0]
subject = [1, 1, 0, 1]
activity = [0, 0, 0, 1]
filer = [0, 1, 1, 0]

Now, I'm ziping them together to get a list of boolean values indicating if any of them are True.

ny_nexus = [True if sum(x) > 0 else False for x in zip(narrative, subject, activity, filer)]

The problem I'm having now, is getting a second list of tuples for which the names of the variables is returned if it had a 1 during the iteration. I imagine it would look something like this:

variables = ("narrative", "subject", "activity", "filer")
reason = [", ".join([some code to filter a tuple]) for x in zip(narrative, subject, activity, filer)]

I just can't figure out how I'd go about this. My desired output would look like this:

reason
# ["subject", "subject, filer", "filer", "subject, activity"]

I'm somewhat new to Python, so I apologize if the solution is easy.

Improve this question
3
  • 1
    By the way, you can say ny_nexus = [sum(x) > 0 for x in zip...] Commented Feb 22, 2016 at 16:47
  • Always appreciate more concise code. Thanks Commented Feb 22, 2016 at 16:48
  • 1
    Even better, use the any() built-in function ;) any([0, 0, 0]) == False, any([0, 1, 0]) == True. So, ny_nexus = [any(x) for x in zip...] Commented Feb 22, 2016 at 17:21

4 Answers 4

Reset to default
3

Store tuples in a dictionary for a cleaner solution:

tups = {'narrative': narrative,
        'subject': subject,
        'activity': activity,
        'filer': filer}

The solution:

reason = [', '.join(k for k, b in zip(tups, x) if b) for x in zip(*tups.values())]

It can also be written using itertools.compress:

from itertools import compress
reason = [', '.join(compress(tups, x)) for x in zip(*tups.values())]

Solutions above do not preserve the order of tuples, e.g. they can return something like

['subject', 'filer, subject', 'filer', 'activity, subject']

If you need the order to be preserved, use collections.OrderedDict as shown below:

from collections import OrderedDict

tups = OrderedDict([
    ('narrative', narrative),
    ('subject', subject),
    ('activity', activity),
    ('filer', filer)
])

# The result is ['subject', 'subject, filer', 'filer', 'subject, activity']

EDIT: The solution that doesn't involve dictionaries:

from itertools import compress
reason = [', '.join(compress(variables, x))
          for x in zip(narrative, subject, activity, filer)]

Consider using dictionaries if the zip(...) call no longer fits on one line.

Improve this answer
Sign up to request clarification or add additional context in comments.

7 Comments

Thus far, this is the only solution that is working. However, it is the one I understand the least. Can you explain what compress does, along with what the role * plays in zip(*tups.values()) please?
@vaultah: I don't see how a dictionary or OrderedDict makes this any cleaner. You've already got compress(), why not just feed it OP's variables tuple?
@brittenb: If you look at the documentation for compress(), you will see that it does pretty much exactly what you're looking for. It selects elements of one sequence based on whether the corresponding elements of another sequence are true.
@JohnY Yeah, I just pulled up the documentation for it and it's pretty self explanatory. Based on that it seems the right solution to this problem is reason = [", ".join(compress(variables, x)) for x in zip(narrative, subject, activity, filer)]
@brittenb: Exactly right. No need in this case to mess with dictionaries or the asterisk operator. (It will be handy to know about, and the official tutorial goes over it, but you can worry about it later.)
|
1

Using zip(narrative, subject, activity, filer) basically transposes the matrix (your list of lists of equal length make up the matrix). You then enumerate through these to find the location n of where the flag is true and index the appropriate variable.

narrative = [0, 0, 0, 0]
subject = [1, 1, 0, 1]
activity = [0, 0, 0, 1]
filer = [0, 1, 1, 0]
variables = ("narrative", "subject", "activity", "filer")
# ========================================================

new_list = [[variables[n] for n, flag in enumerate(indicators) if flag] 
            for indicators in zip(narrative, subject, activity, filer)]
>>> new_list
[['subject'], ['subject', 'filer'], ['filer'], ['subject', 'activity']]

To see the transpose:

>>> [i for i in zip(narrative, subject, activity, filer)]
Improve this answer

Comments

0

You can just use the filtering aspect of the comprehension syntax to get your vaiable English name only if the respective flag is True:

variables = ("narrative", "subject", "activity", "filer")
[tuple (name for flag, name in zip(x, variables) if x)  for x in zip(narrative, subject, activity, filer)]

That said, there is something fishy about your approach - you'd probbly be (far) better of with an object oriented approach there, instead of trying to manually coordinate independent sequences of variables for each of your subjects.

Improve this answer

1 Comment

This approach makes sense to me, but since x will always evaluate as Truthy, it returns all values in variables. Can you elaborate on your object-oriented approach? I'm not married to the way I'm currently doing it and am always willing to try new methods.
0 
    narrative = [0, 0, 0, 0]
    subject = [1, 1, 0, 1]
    activity = [0, 0, 0, 1]
    filer = [0, 1, 1, 0]
    variables = ("narrative", "subject", "activity", "filer")
    ny_nexus = [True if sum(x) > 0 else False for x in zip(narrative, subject, activity, filer)]
    output = []
    [[output.append(variables[j]) if t==1 else None for j,t in enumerate(x)] for x in zip(narrative, subject, activity, filer)]
    print ny_nexus
    print output

Of course you could just do the following without using list comprehensions:

    narrative = [0, 0, 0, 0]
    subject = [1, 1, 0, 1]
    activity = [0, 0, 0, 1]
    filer = [0, 1, 1, 0]
    variables = ("narrative", "subject", "activity", "filer")
    ny_nexus = [True if sum(x) > 0 else False for x in zip(narrative, subject, activity, filer)]
    output = []
    for x in zip(narrative, subject, activity, filer):
        for j,t in enumerate(x):
            output.append(variables[j])
    print ny_nexus
    print output
Improve this answer

3 Comments

This approach also makes sense to me, but it is returning None for all values. Any idea why?
The results of ny_nexus in the above code are actually useless. The useful results are only in output. In fact, you could change the name of ny_nexus to tmp or something and disregard it. The above code would be required in addition to your following statement: ny_nexus = [True if sum(x) > 0 else False for x in zip(narrative, subject, activity, filer)]
Updated answer to reflect

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.