Delete row and column from pandas dataframe

user_10416035 0
Viewed 267

I have a CSV file which is contains a symmetric adjacency matrix which means row and column have equivalent labels.

I would like to import this into a pandas dataframe, ideally have some GUI pop up and ask for a list of items to delete....and then take that list in and set the values in the relative row and column as zero's and return a separate altered dataframe.

In short, something that takes the following matrix

a b c d e

a 0 3 5 3 5

b 3 0 2 4 5

c 5 2 0 1 7

d 3 4 1 0 9

e 5 5 7 9 0

Pops up a simple interface asking "which regions should be deleted" and a line to enter those regions

and say c and e are entered

returns

a b c d e

a 0 3 0 3 0

b 3 0 0 4 0

c 0 0 0 0 0

d 3 4 0 0 0

e 0 0 0 0 0

with the altered entries as shown in bold

it should be able to do this for as many areas as entered which can be up to 379....ideally seperated by commas

pandas user-interface python
2 Answers

Set columns and rows by index values with DataFrame.loc:

vals = ['c','e']
df.loc[vals, :] = 0

df[vals] = 0
#alternative
#df.loc[:, vals] = 0
print (df)
   a  b  c  d  e
a  0  3  0  3  0
b  3  0  0  4  0
c  0  0  0  0  0
d  3  4  0  0  0
e  0  0  0  0  0

Another solution is create boolean mask with numpy broadcasting and set values by DataFrame.mask:

mask = df.index.isin(vals) | df.columns.isin(vals)[:, None]
df = df.mask(mask, 0)
print (df)
   a  b  c  d  e
a  0  3  0  3  0
b  3  0  0  4  0
c  0  0  0  0  0
d  3  4  0  0  0
e  0  0  0  0  0
user_29010020

Start by importing the csv:

import pandas as pd
adj_matrix = pd.read_csv("file/name/to/your.csv", index_col=0)

Then request the input:

regions = input("Please enter the regions that you want deleted (as an array of strings)")
adj_matrix.loc[regions, :] = 0
adj_matrix.loc[:, regions] = 0

Now adj_matrix should be in the form you want.

user_32515400
Related