Pandas - 合并两列时的数据透视表

Question

I have a table with two categorical columns and several numerical ones. I wish to pivot the table on a certain index (in the example the index is the client) while merging the two categorical columns into the names of the numerical ones. Is there a way to do this without a loop?

This is an example table :

Client	Project	Case	Spending	value_mean	value_std
MisterA	Project_15	4	100122	9655.566667	12.70499327
MisterA	Project_15	2	4015	9653	17.66352173
MisterA	Project_15	7	8953	94.4	2.065591118
MisterA	Project_16	1	3922	65519.4	0.894427191
MisterA	Project_16	6	8953	21093.5	17816.03006
MisterA	Project_16	7	8953	30665.3	30643.27374
MisterA	Project_16	2	4015	65517.8	1.788854382
MisterA	Project_16	4	100122	65518.86667	1.153402392
MisterA	Project_16	5	3109	65519	1.632993162
MisterA	Project_18	4	100122	78.84444444	16.89884719
MisterA	Project_18	5	3109	5820	6735.594059
MisterB	Project_15	7	9063	94.6	1.646545205
MisterB	Project_15	2	4015	9636	14.38749457
MisterB	Project_15	6	8968	93.6	1.264911064
MisterB	Project_16	5	4016	65519	1.414213562
MisterB	Project_16	6	8968	22375.3	16701.95844
MisterB	Project_16	7	9063	36482.5	31091.74401
MisterB	Project_16	4	98966	65518.78	1.133333333
MisterB	Project_18	1	2906	79.5	1.914854216
MisterB	Project_18	5	4016	6257	6399.977109
MisterB	Project_18	6	8968	13304.3	52.38330947
MisterB	Project_18	2	4015	78.8	1.095445115

I want it something like this:

Client	Project_15_Case_4_Spending	Project_15_Case_4_value_mean	Project_15_Case_4_value_std	Project_15_Case_2_Spending	Project_15_Case_2_value_mean	Project_15_Case_2_value_std	...
MisterA	100122	9655.566667	12.70499327	4015	9653	17.66352173	...
MisterB	9063	94.6	1.646545205	4015	9636	14.38749457	...

Thank you for you help.

Answer 1

One way -> use pivot_table and then rename the columns . You can use fill_Value=0 if required.

df1 = df.pivot_table(index='Client', columns=['Project', 'Case'], values=[
                     'Spending', 'value_mean', 'value_std'])
df1.columns = [f'{j}_Case_{k}_{i}' for i, j, k in df1.columns]

Answer 2

Use .stack() + .unstack() : you can then keep the column sequence following original row by row 3 columns sequence, as shown in the column sequence in expected output:

df2 = df.set_index(['Client', 'Project', 'Case']).stack().unstack([1,2]).unstack()
df2.columns = df2.columns.map(lambda x: f'{x[0]}_{x[1]}_{x[2]}')
df2 = df2.reset_index()

Result:

print(df2)


    Client  Project_15_4_Spending  Project_15_4_value_mean  Project_15_4_value_std  Project_15_2_Spending  Project_15_2_value_mean  Project_15_2_value_std  Project_15_7_Spending  Project_15_7_value_mean  Project_15_7_value_std  Project_16_1_Spending  Project_16_1_value_mean  Project_16_1_value_std  Project_16_6_Spending  Project_16_6_value_mean  Project_16_6_value_std  Project_16_7_Spending  Project_16_7_value_mean  Project_16_7_value_std  Project_16_2_Spending  Project_16_2_value_mean  Project_16_2_value_std  Project_16_4_Spending  Project_16_4_value_mean  Project_16_4_value_std  Project_16_5_Spending  Project_16_5_value_mean  Project_16_5_value_std  Project_18_4_Spending  Project_18_4_value_mean  Project_18_4_value_std  Project_18_5_Spending  Project_18_5_value_mean  Project_18_5_value_std  Project_15_6_Spending  Project_15_6_value_mean  Project_15_6_value_std  Project_18_1_Spending  Project_18_1_value_mean  Project_18_1_value_std  Project_18_6_Spending  Project_18_6_value_mean  Project_18_6_value_std  Project_18_2_Spending  Project_18_2_value_mean  Project_18_2_value_std
0  MisterA               100122.0              9655.566667               12.704993                 4015.0                   9653.0               17.663522                 8953.0                     94.4                2.065591                 3922.0                  65519.4                0.894427                 8953.0                  21093.5             17816.03006                 8953.0                  30665.3             30643.27374                 4015.0                  65517.8                1.788854               100122.0              65518.86667                1.153402                 3109.0                  65519.0                1.632993               100122.0                78.844444               16.898847                 3109.0                   5820.0             6735.594059                    NaN                      NaN                     NaN                    NaN                      NaN                     NaN                    NaN                      NaN                     NaN                    NaN                      NaN                     NaN
1  MisterB                    NaN                      NaN                     NaN                 4015.0                   9636.0               14.387495                 9063.0                     94.6                1.646545                    NaN                      NaN                     NaN                 8968.0                  22375.3             16701.95844                 9063.0                  36482.5             31091.74401                    NaN                      NaN                     NaN                98966.0              65518.78000                1.133333                 4016.0                  65519.0                1.414214                    NaN                      NaN                     NaN                 4016.0                   6257.0             6399.977109                 8968.0                     93.6                1.264911                 2906.0                     79.5                1.914854                 8968.0                  13304.3               52.383309                 4015.0                     78.8                1.095445