I have three dataframes df_Male , df_female , Df_TransGender
sample dataframe
df_Male
continent avg_count_country avg_age
Asia 55 5
Africa 65 10
Europe 75 8
df_Female
continent avg_count_country avg_age
Asia 50 7
Africa 60 12
Europe 70 0
df_Transgender
continent avg_count_country avg_age
Asia 30 6
Africa 40 11
America 80 10
Now I am concatenating like this below
frames = [df_Male, df_Female, df_Transgender]
df = pd.concat(frames, keys=['Male', 'Female', 'Transgender'])
As you can see America
is present in df_transgender
, same wise Europe is present in df_Male
and df_Female
So I have to concat it in a way so that it looks like below but not manual as there can be huge number of rows
continent avg_count_country avg_age
Male 0 Asia 55 5
1 Africa 65 10
2 Europe 75 8
3 America 0 0
Female 0 Asia 50 7
1 Africa 60 12
2 Europe 70 0
3 America 0 0
Transgender 0 Asia 30 6
1 Africa 40 11
2 America 80 10
3 Europe 0 0
So for other continent
values avg_count_country
and avg_age
should be 0
You can add a "Gender" column before concatenating.
We use Categorical Data with groupby
to calculate the Cartesian product. This should also yield performance benefits.
df = pd.concat([df_Male.assign(gender='Male'),
df_Female.assign(gender='Female'),
df_Transgender.assign(gender='Transgender')])
for col in ['gender', 'continent']:
df[col] = df[col].astype('category')
res = df.groupby(['gender', 'continent']).first().fillna(0).astype(int)
print(res)
avg_count_country avg_age
gender continent
Female Africa 60 12
America 0 0
Asia 50 7
Europe 70 0
Male Africa 65 10
America 0 0
Asia 55 5
Europe 75 8
Transgender Africa 40 11
America 80 10
Asia 30 6
Europe 0 0
You can reindex a bit.
from itertools import product
# Get rid of that number in the index, not sure why you'd need it
df.index = df.index.droplevel(-1)
# Add continents to the index
df = df.set_index('continent', append=True)
# Determine product of indices
ids = list(product(df.index.get_level_values(0).unique(), df.index.get_level_values(1).unique()))
# Reindex and fill missing with 0
df = df.reindex(ids).fillna(0).reset_index(level=-1)
df
is now:
continent avg_count_country avg_age
Male Asia 55.0 5.0
Male Africa 65.0 10.0
Male Europe 75.0 8.0
Male America 0.0 0.0
Female Asia 50.0 7.0
Female Africa 60.0 12.0
Female Europe 70.0 0.0
Female America 0.0 0.0
Transgender Asia 30.0 6.0
Transgender Africa 40.0 11.0
Transgender Europe 0.0 0.0
Transgender America 80.0 10.0
If you want that other numeric index, then you can just do: df.groupby(df.index).cumcount()
to number the values in each group.
Making use of DataFrame.pivot
, a slight modification to @jpp's answer allows you to avoid having to manually manipulate indices:
df = pd.concat([df_Male.assign(gender='Male'),
df_Female.assign(gender='Female'),
df_Transgender.assign(gender='Transgender')])
df.pivot('gender', 'continent').fillna(0).stack().astype(int)
avg_count_country avg_age
gender continent
Female Africa 60 12
America 0 0
Asia 50 7
Europe 70 0
Male Africa 65 10
America 0 0
Asia 55 5
Europe 75 8
Transgender Africa 40 11
America 80 10
Asia 30 6
Europe 0 0
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.