Merge two arrays into an list of arrays

Question

Hi I am trying compile a bunch of arrays that I have in a dictionary using a for loop and I cant seem to find proper solution for this.

Essentially what I have in a simplified form:

dict['w1']=[1,2,3]
dict['w2']=[4,5,6]
dict['w3']=[7,8]

x = []
for i in range(3):
    x = np.concatenate([x],[dict['w'+str(i+1)].values],axis=0)

what it gives:

x = [1,2,3,4,5,6,7,8]

what I want:

x = [[1,2,3],[4,5,6],[6,7]]

I want to use the for loop because I have too many arrays to 'compile' and cant key in them one by one would be very inefficient. This way I can use the created array to plot a boxplot directly.

A simiiliar question exists with no loop requirement but still does not have proper solution. Link

You won't be able to achieve that solution with a numpy array, all column and rows have to be the same size in a np.array. That being said you seem to be confusing numpy arrays and python lists they are not the same, so don't use numpy functions (wich work on arrays) if you only use lists. — jadsq
– jadsq, Commented Oct 8, 2016 at 12:41
Last element of expected o/p has to be [7,8] instead, right? — Divakar
– Divakar, Commented Oct 8, 2016 at 12:58
Also, I sensed that you are looking to avoid loop from your question's other parts, except at this line : "I want to use the for loop ...". Typo there? — Divakar
– Divakar, Commented Oct 8, 2016 at 13:07
I'm removing the duplicate; the answers in stackoverflow.com/questions/9285414/appending-to-a-nested-list are inferior to the answers already given (and accepted) here. — hpaulj
– hpaulj, Commented Oct 8, 2016 at 18:22

LycuiD · Accepted Answer · 2016-10-08 12:51:36Z

2

Comprehension:

# using name as "dict" is not proper use "_dict" otherwise.
dict['w1']=[1,2,3]
dict['w2']=[4,5,6]
dict['w3']=[7,8]

x = [dict["w"+str(i)] for i in range(1, 4)]

gives output:

[[1, 2, 3], [4, 5, 6], [7, 8]]

edited Oct 8, 2016 at 12:51

answered Oct 8, 2016 at 12:45

LycuiD

2,5751 gold badge21 silver badges28 bronze badges

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

2 Comments

xplodnow Over a year ago

This works like a charm! I have a question though. I have the keys in the dict as 'w01', 'w02'.......'w10'. So there is slight issue when its hits the value '10'. Is it possible to include a 'if' into the code u have written above?

LycuiD Over a year ago

You can use condition in Comprehensions, it'll go something like [x1 for x1 in something if condition == True else x2].

Divakar · Accepted Answer · 2016-10-08 12:43:46Z

1

One approach would be to get the argsort indices based on the keys and sort the elements/values out of the dictionary using those indices, like so -

np.take(dict.values(),np.argsort(dict.keys()))

If you need a list as the output, add .tolist().

Sample run -

In [84]: dict
Out[84]: {'w1': [4, 8, 2, 1], 'w2': [1, 3, 8], 'w3': [3, 2]}

In [85]: np.take(dict.values(),np.argsort(dict.keys()))
Out[85]: array([[4, 8, 2, 1], [1, 3, 8], [3, 2]], dtype=object)

In [86]: np.take(dict.values(),np.argsort(dict.keys())).tolist()
Out[86]: [[4, 8, 2, 1], [1, 3, 8], [3, 2]]

answered Oct 8, 2016 at 12:43

Divakar

222k19 gold badges273 silver badges374 bronze badges

Comments

Mureinik · Accepted Answer · 2016-10-08 12:43:44Z

0

Just append the item to the list. Note, BTW, that range starts with 0 if you don't specify the start value and treats the end value as an exclusive border:

x = []
for i in range(1, 4):
    x.append(dict['w' + str(i)])

answered Oct 8, 2016 at 12:43

Mureinik

316k54 gold badges400 silver badges405 bronze badges

Comments

Jay Holister · Accepted Answer · 2016-10-08 12:50:04Z

0

If you mean lists and not dictionarys, then you're looking for the zip command.

Zip (w1,w2,w3)

answered Oct 8, 2016 at 12:50

Jay Holister

1312 bronze badges

Collectives™ on Stack Overflow

Merge two arrays into an list of arrays

4 Answers 4

2 Comments

Comments

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

2 Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related