divide a dataframe into another dataframe (element by element) - python-polars

Question

In pandas, I could set several named columns as an index and find the quotient of the division of two
DataFrame,like this
import pandas as pd
df_1=pd.DataFrame( {
'Name':['a','a','a', 'b','b', 'c'],
'Name_2':['first','second','third', 'first','second', 'first'],
'Value':[20,40,50,100,150,400]
})
df_2=pd.DataFrame( {
'Name':['a','a','a', 'b','b', 'c'],
'Name_2':['first','second','third', 'first','second', 'first'],
'Value':[10,20,25,50,75,200]
})
df_1=df_1.set_index(['Name','Name_2'])
df_2=df_2.set_index(['Name','Name_2'])
df_1/df_2
How can something like this be implemented in python-polars?
I can't find an answer to this question in the documentation.

Answer 1

You just use a join, then do the math on the appropriate column(s).
df_1=pl.DataFrame( {
'Name':['a','a','a', 'b','b', 'c'],
'Name_2':['first','second','third', 'first','second', 'first'],
'Value':[20,40,50,100,150,400]
})
df_2=pl.DataFrame( {
'Name':['a','a','a', 'b','b', 'c'],
'Name_2':['first','second','third', 'first','second', 'first'],
'Value':[10,20,25,50,75,200]
})
That's the setup then the solution is:
df_1.join(df_2, on=['Name','Name_2']) \
.select(['Name','Name_2', pl.col('Value')/pl.col('Value_right')])
If you have a bunch of "Value" columns and different indx columns you can do something like:
myindxcols=['Name', 'Name_2]
myvalcols=[x for x in df_1.columns if x in df_2.columns and not x in myindxcols]
df_1.join(df_2, on=myindxcols) \
.select(myindxcols + [pl.col(x)/pl.col(f"{x}_right") for x in myvalcols])

Answer 2

#Dean MacGregor beat me to it. Please accept his answer.
df_1 = pl.DataFrame( {
"Name":["a","a","a", "b","b", "c"],
"Name_2":["first","second","third", "first","second", "first"],
"Value":[20,40,50,100,150,400]
})
df_2 = pl.DataFrame( {
"Name":["a","a","a", "b","b", "c"],
"Name_2":["first","second","third", "first","second", "first"],
"Value":[10,20,25,50,75,200]
})
keys = ["Name", "Name_2"]
(df_1
.join(df_2, on=keys, suffix="_right")
.select([
*keys,
pl.col("Value") / pl.col("Value_right")
])
)
shape: (6, 3)
┌──────┬────────┬───────┐
│ Name ┆ Name_2 ┆ Value │
│ --- ┆ --- ┆ --- │
│ str ┆ str ┆ f64 │
╞══════╪════════╪═══════╡
│ a ┆ first ┆ 2.0 │
│ a ┆ second ┆ 2.0 │
│ a ┆ third ┆ 2.0 │
│ b ┆ first ┆ 2.0 │
│ b ┆ second ┆ 2.0 │
│ c ┆ first ┆ 2.0 │
└──────┴────────┴───────┘