Your data is often in a “wide” format, like a spreadsheet, but analysis tools (like plotting libraries) prefer “long” format. One useful function for this transformation is Polars melt, which helps you reshape data efficiently.
Wide Data (Bad for analysis): | student | exam1 | exam2 | | :— | :— | :— | | Alice | 90 | 85 | | Bob | 80 | 95 |
Long Data (Good for analysis): | student | exam | score | | :— | :— | :— | | Alice | exam1 | 90 | | Alice | exam2 | 85 | | Bob | exam1 | 80 | | Bob | exam2 | 95 |
The Polars function to do this transformation is melt().
The Code
import polars as pl
df_wide = pl.DataFrame({
"student": ["Alice", "Bob"],
"exam1": [90, 80],
"exam2": [85, 95]
})
# 1. 'id_vars' are the columns to KEEP (don't unpivot)
# 2. 'value_vars' are the columns to MELT DOWN
df_long = df_wide.melt(
id_vars="student",
value_vars=["exam1", "exam2"],
variable_name="exam", # Name for the new "key" column
value_name="score" # Name for the new "value" column
)
print(df_long)Output:
shape: (4, 3) ┌─────────┬───────┬───────┐ │ student ┆ exam ┆ score │ │ --- ┆ --- ┆ --- │ │ str ┆ str ┆ i64 │ ╞═════════╪═══════╪═══════╡ │ Alice ┆ exam1 ┆ 90 │ │ Bob ┆ exam1 ┆ 80 │ │ Alice ┆ exam2 ┆ 85 │ │ Bob ┆ exam2 ┆ 95 │ └─────────┴───────┴───────┘
This “tidy data” format is essential for plotting with libraries like Seaborn and is a fundamental skill in data cleaning.
