gillesa's questions

我正在将一些代码从 迁移Pandas到Polars。我尝试使用cut但polars存在差异（没有bin，所以我必须计算它）。

label但我还是不明白极坐标的结果。

我必须使用比我想要的更多的标签才能获得相同的结果pandas。

import numpy as np
import pandas as pd
import polars as pl

# Exemple de DataFrame Polars
data = {
    "value": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
}
df_pl = pl.DataFrame(data)

# Convertir en DataFrame Pandas pour obtenir les breakpoints
df_pd = df_pl.to_pandas()

# Use returbins to get the breakpoints (from pandas)
df_pd["cut_label_pd"], breakpoints = pd.cut(df_pd["value"], 4, labels=["low", "medium", "hight", "very high"], retbins=True)
print(pl.from_pandas(df_pd))

shape: (10, 2)
┌───────┬──────────────┐
│ value ┆ cut_label_pd │
│ ---   ┆ ---          │
│ i64   ┆ cat          │
╞═══════╪══════════════╡
│ 1     ┆ low          │
│ 2     ┆ low          │
│ 3     ┆ low          │
│ 4     ┆ medium       │
│ 5     ┆ medium       │
│ 6     ┆ hight        │
│ 7     ┆ hight        │
│ 8     ┆ very high    │
│ 9     ┆ very high    │
│ 10    ┆ very high    │
└───────┴──────────────┘

print(breakpoints)
# [ 0.991  3.25   5.5    7.75  10.   ]

labels有没有更好的方法？（注意中的值polars cut）

# Cut in polars
labels = ["don't use it", "low", "medium", "hight", "very high", "don't use it too"] 
df_pl = df_pl.with_columns(
    pl.col("value").cut(breaks=breakpoints, labels=labels).alias("cut_label_pl")
)

print(df_pl)

shape: (10, 2)
┌───────┬──────────────┐
│ value ┆ cut_label_pl │
│ ---   ┆ ---          │
│ i64   ┆ cat          │
╞═══════╪══════════════╡
│ 1     ┆ low          │
│ 2     ┆ low          │
│ 3     ┆ low          │
│ 4     ┆ medium       │
│ 5     ┆ medium       │
│ 6     ┆ hight        │
│ 7     ┆ hight        │
│ 8     ┆ very high    │
│ 9     ┆ very high    │
│ 10    ┆ very high    │
└───────┴──────────────┘