Python: Create lists from diagonal values in a dataframe

Question

I'm trying to create a routine in Python to collect every diagonal group of values in df. Here's a reproducible example of what I'm trying to achieve:

data = {'column1':[1,1, 2, 3,6, 4,5,6], 'column2':[np.nan,4,3,5,6,2,3,4], 'column3':[np.nan,np.nan,3,2,5,np.nan,8,4], 'column4':[np.nan,np.nan,np.nan,3,6,np.nan,np.nan, 6], 'column5':[np.nan, np.nan, np.nan, np.nan, 8, np.nan, np.nan,np.nan]}

df = pd.DataFrame(data, columns = ['column1', 'column2', 'column3', 'column4', 'column5'])
my_list = []
# dict_list = {'list' + str(i):[] for i in list(range(len(df)))}

for i in range(len(df)):
    for j in range(len(df.columns)):   
        
        if (i + j) < df.iloc[6,2]:
            my_list.append(df.iloc[i + j, j])
            
        else:
            break

This code returns me one single list:

my_list = [1,4.0,3.0,3.0,8.0,1,3.0,2.0,6.0,nan,2,5.0,5.0,nan,nan,3,6.0,nan,nan,nan,6,2.0,8.0,6.0,4,3.0,40,5,4.0,6]

And based on the structure of the given df, what I'm trying to achieve is:

dict_list = [[1,4,3,3,8],[1,3,2,6],[2,5,5],[3,6],[6,2,8,6],[4,3,4],[5,4],[6]]

From what I've seen I could do this by creating a list of lists (commented in the code as dict_list, here's the reference: Python : creating multiple lists), but I haven't been able to put my data as shown in dict_listobject.

I will appreciate any help or guide.

Thank you!

Yes it solved it, I've already answered to you too and accepted the solution as the one I was looking for. Cheers! — XioHu, Aug 14 '20 at 21:42

Youness Saadna · Accepted Answer · 2020-08-14T20:27:22.240

Using the numpy.diag() will help you

This is the code I used:

import pandas as pd
import numpy as np

data = {'column1':[1,1, 2, 3,6, 4,5,6], 'column2':[np.nan,4,3,5,6,2,3,4], 'column3':[np.nan,np.nan,3,2,5,np.nan,8,4], 'column4':[np.nan,np.nan,np.nan,3,6,np.nan,np.nan, 6], 'column5':[np.nan, np.nan, np.nan, np.nan, 8, np.nan, np.nan,np.nan]}
df = pd.DataFrame(data, columns = ['column1', 'column2', 'column3', 'column4', 'column5'])
nump=df.to_numpy()

my_list = []
for i in range(len(nump)):
    my_list.append(np.diag(nump,k=-(i)))

OUTPUT:

[array([1., 4., 3., 3., 8.]),
 array([ 1.,  3.,  2.,  6., nan]),
 array([ 2.,  5.,  5., nan, nan]),
 array([ 3.,  6., nan, nan, nan]),
 array([6., 2., 8., 6.]),
 array([4., 3., 4.]),
 array([5., 4.]),
 array([6.])]

To clean nan values:

cleanedList=[]

for i in range(len(my_list)):
    l=[x for x in my_list[i] if str(x) != 'nan']
    print(l)
    cleanedList.append(l)

OUTPUT:

[[1.0, 4.0, 3.0, 3.0, 8.0],
 [1.0, 3.0, 2.0, 6.0],
 [2.0, 5.0, 5.0],
 [3.0, 6.0],
 [6.0, 2.0, 8.0, 6.0],
 [4.0, 3.0, 4.0],
 [5.0, 4.0],
 [6.0]]

For more information about how to use numpy.diag() visit the documentation numpy.diag

Thank you very much @Youness Saadna, your solution fitted exactly what I was looking for! Besides, Numpy allows you to work incredible faster when doing for loops. :) — XioHu, Aug 14 '20 at 21:40

Python: Create lists from diagonal values in a dataframe

1 Answers1