how to use pd.cut() across columns of a data frame?

Question

>> df = pd.DataFrame(np.random.rand(10,4))
>> pd.cut(df,[0,0.5,1])

ValueError: Input array must be 1 dimensional

How can I get pd.cut() to work across all columns of a data frame?

Answer 1

Use apply

df.apply(pd.cut, bins=[0,0.5,1])

You can specify the axis if you want to run across columns (axis=0) or rows (axis=1)

Answer 2

If you don't mind a slightly different type of labeling, numpy.digitize provides a vectorized n-d solution.

---

np.digitize(df, bins=[0, 0.5, 1.0])

array([[2, 2, 2, 2],
       [1, 2, 2, 2],
       [1, 1, 2, 1],
       [2, 1, 2, 1],
       [2, 1, 2, 1],
       [2, 2, 2, 2],
       [1, 2, 1, 1],
       [2, 1, 2, 2],
       [2, 2, 1, 1],
       [2, 1, 2, 1]], dtype=int64)

---

The label 1 would correspond to 0-0.5, 2 to 0.5-1.0, etc.

---

Performance

df = pd.DataFrame(np.random.rand(1000, 1000))

%timeit pd.DataFrame(np.digitize(df, bins=[0, 0.5, 1.0]), columns=df.columns)
13.2 ms ± 36.8 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

%timeit df.apply(pd.cut, bins=[0, 0.5, 1])
3.11 s ± 12.4 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

%timeit pd.cut(df.stack(),[0,0.5,1]).unstack()
1.48 s ± 3.82 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)