I have just started learning ‘Polars’ and trying to convert my pandas code to polars.

Here is my code:

import polars as pl
import pandas as pd

pdf = pl.DataFrame({
    "Country": ["US", "UK", "US"],
    "Date": pd.to_datetime(["2023-01-01", "2023-02-01", "2023-03-01"])
})

df = pdf.to_pandas()

# Using Pandas
out = df.groupby('Country').apply(lambda row: row['Date'].dt.strftime('%Y-%m').value_counts()).unstack(0).sort_index()
print(out)

# How to do that in polars? (my attempt so far)
pdf = pdf.with_columns(
    Date_ym = pdf['Date'].dt.strftime("%Y-%m")
)


out = (pdf.groupby("Country")
 .agg(pl.col('Date_ym')
      .value_counts())
 .explode('Date_ym')
)

print(out)

Required output

Country UK US
2023-01 NaN 1.0
2023-02 1.0 NaN
2023-03 NaN 1.0

>Solution :

It looks like a .pivot()

(df.pivot('Date', 'Date', 'Country', aggregate_function='count')
   .with_columns(
      pl.col('Date').dt.to_string('%Y-%m')
   )
)   

shape: (3, 3)
┌─────────┬──────┬──────┐
│ Date    ┆ US   ┆ UK   │
│ ---     ┆ ---  ┆ ---  │
│ str     ┆ u32  ┆ u32  │
╞═════════╪══════╪══════╡
│ 2023-01 ┆ 1    ┆ null │
│ 2023-02 ┆ null ┆ 1    │
│ 2023-03 ┆ 1    ┆ null │
└─────────┴──────┴──────┘