对 Pivot 表的列进行排序 Pandas

Question

I'm trying to sort the columns of a pivot table in a specific order, not alphabetical. In my table, I have the same values on the x and y axis and the values are the similarity between them. I have the order of the y axis sorted in a specific order and would like the x axis of values to match, so that the diagonal of values will be all 1.0. Below is an example of what I'm doing, with example data

# Code to create pivot table
pivot=final_df.groupby(['Fruit A','Fruit B'], sort=False)['Similarity Score'].sum().unstack('Fruit B')

	Apple	Orange	Mango	Banana
Orange	0.4	1	0.6	0.7
Mango	0.3	0.4	1	0.5
Apple	1	0.3	0.6	0.1
Banana	0.4	0.2	0.5	1

Ideally, I want the above table to also say orange, mango, apple, banana (in that order) on the column labels, resulting in 1 throughout the diagonal. How can I do this effectively?

Answer 1

One option: use an ordered Categorical for your two fruits columns.

order = ['Orange', 'Mango', 'Apple', 'Banana']
df['Fruit A'] = pd.Categorical(df['Fruit A'], categories=order, ordered=True)
df['Fruit B'] = pd.Categorical(df['Fruit B'], categories=order, ordered=True)
# then pivot

Second option: reindex :

order = ['Orange', 'Mango', 'Apple', 'Banana']
pivot.reindex(index=order, columns=order)