I need help creating a Python function to achieve the following:
1) Take 3 Pandas dataframes as input (containing an index column, and an associated integer or float value in the second column). These are defined as follows:
import pandas as pd
df1=pd.DataFrame([['placementA',2],['placementB',4]],columns=
['placement','value'])
df1.set_index('placement',inplace=True)
df2=pd.DataFrame([['strategyA',1],['strategyB',5],['strategyC',6]],columns=
['strategy','value'])
df2.set_index('strategy',inplace=True)
df3=pd.DataFrame([['categoryA',1.5],['categoryB',2.5]],columns=
['category','value'])
df3.set_index('category',inplace=True)
2) Using these three dataframes, create a new dataframe ('df4') which organises all possible combinations of the 3 indices across the first 3 columns;
3) In the 4th column, append the mathematical product of all the associated 'values' from the three source dataframes. The DataFrame output of the function should therefore look like: https://ibb.co/cypEY6
Many thanks in advance for your help.
Colin
Use product
of all indexes and columns and create DataFrame
by constructor, for multiple all columns use prod
:
from itertools import product
names = ['placement','strategy','category']
mux = pd.MultiIndex.from_product([df1.index, df2.index, df3.index], names=names)
df = (pd.DataFrame(list(product(df1['value'], df2['value'], df3['value'])), index=mux)
.prod(1).reset_index(name='mult'))
print (df)
placement strategy category mult
0 placementA strategyA categoryA 3.0
1 placementA strategyA categoryB 5.0
2 placementA strategyB categoryA 15.0
3 placementA strategyB categoryB 25.0
4 placementA strategyC categoryA 18.0
5 placementA strategyC categoryB 30.0
6 placementB strategyA categoryA 6.0
7 placementB strategyA categoryB 10.0
8 placementB strategyB categoryA 30.0
9 placementB strategyB categoryB 50.0
10 placementB strategyC categoryA 36.0
11 placementB strategyC categoryB 60.0
Alternative is multiple
all values by list comprehension:
import operator
import functools
from itertools import product
names = ['placement','strategy','category']
a = list(product(df1.index, df2.index, df3.index))
b = product(df1['value'], df2['value'], df3['value'])
data = [functools.reduce(operator.mul, x, 1) for x in b]
df = pd.DataFrame(a, columns=names).assign(mult=data)
print (df)
placement strategy category mult
0 placementA strategyA categoryA 3.0
1 placementA strategyA categoryB 5.0
2 placementA strategyB categoryA 15.0
3 placementA strategyB categoryB 25.0
4 placementA strategyC categoryA 18.0
5 placementA strategyC categoryB 30.0
6 placementB strategyA categoryA 6.0
7 placementB strategyA categoryB 10.0
8 placementB strategyB categoryA 30.0
9 placementB strategyB categoryB 50.0
10 placementB strategyC categoryA 36.0
11 placementB strategyC categoryB 60.0
Dynamic solution with list of DataFrames
, only is necessary same columnname value
in each one:
dfs = [df1, df2, df3]
names = ['placement','strategy','category']
a = list(product(*[x.index for x in dfs]))
b = list(product(*[x['value'] for x in dfs]))
data = pd.DataFrame(b).product(1)
df = pd.DataFrame(a, columns=names).assign(mult=data)
print (df)
placement strategy category mult
0 placementA strategyA categoryA 3.0
1 placementA strategyA categoryB 5.0
2 placementA strategyB categoryA 15.0
3 placementA strategyB categoryB 25.0
4 placementA strategyC categoryA 18.0
5 placementA strategyC categoryB 30.0
6 placementB strategyA categoryA 6.0
7 placementB strategyA categoryB 10.0
8 placementB strategyB categoryA 30.0
9 placementB strategyB categoryB 50.0
10 placementB strategyC categoryA 36.0
11 placementB strategyC categoryB 60.0
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.