Convert dataframe data from columns to rows

Question

In Pandas, I'm trying to relocate data placed in several columns, to several rows in the sense suggested by the two tables below.

Each customer should appear in as many rows as they made purchases. If it matters, column j+1 can only have a value if column j does too.

How do I go about this?

This is called [**converting from wide-form to long-form**](https://stackoverflow.com/search?q=%5Bpython%5D+long+form+wide+form) (or "melting"). There are many duplicates already, please pick one of those 49 hits and close-as-duplicate. — smci, Jan 23 '18 at 18:46
@smci Thank you very much for the terminology. That's useful. I will do as you suggest. — Git Gud, Jan 23 '18 at 18:47

score 1 · Accepted Answer · answered Jan 23 '18 at 18:41

Pandas offers something specifically meant for this called melt: https://pandas.pydata.org/pandas-docs/stable/generated/pandas.melt.html

Without a sample of your code I can't give an example that is code specific, but you would have something like this:

id_vars = ['Customer']
var_name = 'Purchase Type'
value_name = 'Value'

melted_df = pf.melt(unmelted_df, id_vars=id_vars, var_name=var_name, value_name=value_name)

As a result, you would get a melted DataFrame where Purchase 1, Purchase 2, etc are attributes of 'Purchase Type' and the value of those original purchases is displayed in Value.

This was enough. Thank you very much. – Git Gud Jan 23 '18 at 18:51 — Git Gud, Jan 23 '18 at 18:51

Convert dataframe data from columns to rows

1 Answers1