Pandas pivot table and merge column names

user_9649473 0
Viewed 72

Here's the thing,

I need to pivot a table that is in the long format to its wide format.

Example:

d = {'col1': ['a','b', 'c', 'b', 'c', 'c'], 'col2': ['joe', 'mary', 'mary', 'joseph', 'allan', 'mary'], 
'col3':[2020, 2022, 2019, 2020, 2022, 2019], 'values':[123,235,3456,213,423,52]}
df = pd.DataFrame(data=d)
print(df)

I would need something like

df = df.pivot(index=["col1"], columns=["col3", "col2"], values="values")

For generating this structure with values if exists and null if doesnt:

col1;2019_joe;2019_joe;2019_joe;2019_mary;2020_mary;2021_mary
a
b
c
d
e
f

But the command pivot() doesnt accept a list on column keyword argument.

pandas pivot python
3 Answers

Problem is that you have two same c mary 2019, in this case, you can try pivot_table with aggfunc argument

out = df.pivot_table(index=["col1"], columns=["col3", "col2"], values="values", aggfunc='mean')
out.columns = out.columns.map(lambda index: f'{index[0]}_{index[1]}')

print(out)

      2019_mary  2020_joe  2020_joseph  2022_allan  2022_mary
col1
a           NaN     123.0          NaN         NaN        NaN
b           NaN       NaN        213.0         NaN      235.0
c        1754.0       NaN          NaN       423.0        NaN
user_103151630

Does this get you the results you are expecting?

df['col4'] = df['col3'].astype(str) + '_' + df['col2'].astype(str)
df = df.drop(columns = ['col2', 'col3'])
pd.pivot_table(df, index = 'col1', values = 'values', columns = 'col4').reset_index()
user_177504310

I don't get what is the logic to be applied on col1. In your constructor you have 'col1': ['a','b', 'c', 'b', 'c', 'c'] but then in your expected output you have a, b, c, d, e and f. Where are d, e and f coming from? Assuming that you don't care about col1, what about df.pivot(columns=['col3','col2'])['values']?

col3   2020   2022    2019   2020   2022
col2    joe   mary    mary joseph  allan
0     123.0    NaN     NaN    NaN    NaN
1       NaN  235.0     NaN    NaN    NaN
2       NaN    NaN  3456.0    NaN    NaN
3       NaN    NaN     NaN  213.0    NaN
4       NaN    NaN     NaN    NaN  423.0
5       NaN    NaN    52.0    NaN    NaN
user_31900760
Related