如何相互减去所有日期列（以排列方式）并将它们存储在新的 pandas DataFrame 中？

Question

I was working on Jupyter and arrived at a situation where I had to take differences of each column from every other column taken in permutation and then store them in a separate DataFrame. I tried using nested loops but got stuck while assigning the values to the DataFrame.

n=0
for i in range(len(list(df.columns))-1):
  for j in range(i+1, len(list(df.columns))-1):
    df1[n] = pd.DataFrame(abs((df.iloc[:,i] - df.iloc[:,j]).dt.days))
    n=n+1
df1

Also, I would like to have column headers in this format: D1-D2, D1-D3, etc. The difference in dates has to be a positive integer. I would really appreciate if anyone could help me with this code. Thanks!

A snippet of the DataFrame

Answer 1

import itertools
import pandas as pd

# create a sample dataframe
df = pd.DataFrame(data={"co1":[1,2,3,4], "co22":[4,3,2,1], "co3":[2,3,2,4]})

# iterate over all permutations of size 2 and write to dictionary
newcols = {}
for col1, col2 in itertools.permutations(df.columns, 2):
    newcols["-".join([col1, col2])] = df[col1]-df[col2]

# create dataframe from dict
newdf = pd.DataFrame(newcols)