Pivot pandas dataframe 具有多索引列

Question

I am struggling with how to pivot the dataframe to have multi-indexed columns.

I have a dataframe like this:

data = pd.DataFrame({"name":["a", "a", "b", "b", "b", "b",  "c", "c"], 
                     "month":[1, 1, 2, 3, 1, 1, 1, 3], 
                     "buy_sell":["sell", "buy", "sell", "buy", "sell", "buy", "sell", "buy"],
                     "value":[10, 20, 30, 40, 20, 80, 50, 60]})
data

And I want to pivot this to a wide format. index is name , and for columns I want to create multi-index with the combination of month and buy_sell

Or if not multi-indexed columns, I want to pivot the dataframe so that the columns have the suffices such as sell_1 , buy_1 , sell_2 , buy_2 etc...

Any help would be appreciated. Thank you!

Answer 1

You can use set_index() and unstack() :

(data.set_index(['name','month','buy_sell'])['value']
      .unstack(['month','buy_sell']))

Output:

month        1           2     3
buy_sell  sell   buy  sell   buy
name                            
a         10.0  20.0   NaN   NaN
b         20.0  80.0  30.0  40.0
c         50.0   NaN   NaN  60.0

If you insist on having those columns with all NaN , you can unstack one level at a time:

(data.set_index(['name','month','buy_sell'])['value']
      .unstack('month').unstack('buy_sell')
     )

Output:

month        1         2           3     
buy_sell   buy  sell buy  sell   buy sell
name                                     
a         20.0  10.0 NaN   NaN   NaN  NaN
b         80.0  20.0 NaN  30.0  40.0  NaN
c          NaN  50.0 NaN   NaN  60.0  NaN