下面是另一种方法，可以保留每列中nan数小于或等于指定值的列：

max_number_of_nas = 3000
df = df.loc[:, (df.isnull().sum(axis=0) <= max_number_of_nas)]

在我的测试中，这似乎比Jianxun Li在我测试的案例中建议的drop columns方法要快一些（如下所示）。然而，我应该注意，如果你只是简单地不使用apply方法（例如df.drop(df.columns[df.isnull().sum(axis=0) > max_number_of_nans], axis=1)），性能会变得更相似。

np.random.seed(0)
df = pd.DataFrame(np.random.randn(10000,5), columns=list('ABCDE'))
df[df < 0] = np.nan
max_number_of_nans = 5010

%timeit c = df.loc[:, (df.isnull().sum(axis=0) <= max_number_of_nans)]
>> 1.1 ms ± 4.08 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

%timeit c = df.drop(df.columns[df.isnull().sum(axis=0) > max_number_of_nans], axis=1)
>> 1.3 ms ± 11.8 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

%timeit c = df.drop(df.columns[df.apply(lambda col: col.isnull().sum() > max_number_of_nans)], axis=1)
>> 2.11 ms ± 29.4 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

性能通常随数据大小而变化，因此不要忘记检查与数据最接近的情况。

np.random.seed(0)
df = pd.DataFrame(np.random.randn(10, 5), columns=list('ABCDE'))
df[df < 0] = np.nan
max_number_of_nans = 5

%timeit c = df.loc[:, (df.isnull().sum(axis=0) <= max_number_of_nans)]
>> 755 µs ± 4.84 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

%timeit c = df.drop(df.columns[df.isnull().sum(axis=0) > max_number_of_nans], axis=1)
>> 777 µs ± 12 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

%timeit c = df.drop(df.columns[df.apply(lambda col: col.isnull().sum() > max_number_of_nans)], axis=1)
>> 1.71 ms ± 17.3 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)

同样的逻辑，只是把所有的东西放在一行。

import pandas as pd
import numpy as np

# artificial data
# ====================================
np.random.seed(0)
df = pd.DataFrame(np.random.randn(10,5), columns=list('ABCDE'))
df[df < 0] = np.nan

        A       B       C       D       E
0  1.7641  0.4002  0.9787  2.2409  1.8676
1     NaN  0.9501     NaN     NaN  0.4106
2  0.1440  1.4543  0.7610  0.1217  0.4439
3  0.3337  1.4941     NaN  0.3131     NaN
4     NaN  0.6536  0.8644     NaN  2.2698
5     NaN  0.0458     NaN  1.5328  1.4694
6  0.1549  0.3782     NaN     NaN     NaN
7  0.1563  1.2303  1.2024     NaN     NaN
8     NaN     NaN     NaN  1.9508     NaN
9     NaN     NaN  0.7775     NaN     NaN

# processing: drop columns with no. of NaN > 3
# ====================================
df.drop(df.columns[df.apply(lambda col: col.isnull().sum() > 3)], axis=1)

Out[183]:
        B
0  0.4002
1  0.9501
2  1.4543
3  1.4941
4  0.6536
5  0.0458
6  0.3782
7  1.2303
8     NaN
9     NaN

对我来说，我似乎不需要set_index：

df = (df.T
     .loc[lambda x: ((x['label'] > .05) | (x['label'] < -.05))]
     .T.reset_index().set_index('index'))