Downcasting numeric columns

Now that the ranges look safe, cast the numeric columns to smaller dtypes. Use Int32 for the integer columns and Float32 for the floats where lower precision is still good enough for summary stats.

The movies DataFrame is preloaded for you.

Diese Übung ist Teil des Kurses

<Kurs>Scaling and Optimizing Data Pipelines with Polars</Kurs>

Kurs ansehen

Übungsanweisungen

Cast vote_count and budget to pl.Int32.
Cast runtime and vote_average to pl.Float32.

Interaktive praktische Übung

Versuche dich an dieser Übung, indem du diesen Beispielcode vervollständigst.

movies_optimized = movies.with_columns(
    # Integer columns to Int32
    pl.col("vote_count").cast(pl.____),
    pl.col("budget").cast(pl.____),
    # Float columns to Float32
    pl.col("runtime").cast(pl.____),
    pl.col("vote_average").cast(pl.____),
)

result = movies_optimized.select(
    "movie_title", "budget", "runtime", "vote_average", "vote_count"
).head(8)
print(result)

Code bearbeiten und ausführen