[Python] Following and Count of State Changes Between Columns in Polars

I have a dataframe with multiple IDs and corresponding states. I want to analyze how the states have changed over time and present this information effectively.

Here is an example:

import polars as pl

df = pl.DataFrame({
"ID": [1, 2, 3],
"T0": ["A", "B", "C"],
"T1": ["B", "B", "A"],
})


One aproach it's to "concat" the columns , and then do a value_counts() of the Change column

df = df.with_columns(
(pl.col("T0") + " -> " + pl.col("T1")).alias("Change")
)


However, there might be a better approach to this, or even a built-in function that can achieve what I need more efficiently.

Current Output:

shape: (3, 4)
┌─────┬─────┬─────┬────────┐
│ ID ┆ T0 ┆ T1 ┆ Change │
│ --- ┆ --- ┆ --- ┆ --- │
│ i64 ┆ str ┆ str ┆ str │
╞═════╪═════╪═════╪════════╡
│ 1 ┆ A ┆ B ┆ A -> B │
│ 2 ┆ B ┆ B ┆ B -> B │
│ 3 ┆ C ┆ A ┆ C -> A │
└─────┴─────┴─────┴────────┘

shape: (3, 2)
┌────────┬───────┐
│ Change ┆ count │
│ --- ┆ --- │
│ str ┆ u32 │
╞════════╪═══════╡
│ C -> A ┆ 1 │
│ B -> B ┆ 1 │
│ A -> B ┆ 1 │
└────────┴───────┘

Continue reading...