我想把python pandas DataFrame对象保存为pickle。使用pandas.to_pickle和使用pickle.dumps有什么区别？
我做了一些测试。下面是我的测试代码：

import pandas as pd
import pickle
df = pd.DataFrame([[1, 2], [3, 4]], columns=list('AB'), index=['x', 'y'])

# Save
df.to_pickle('df1.pickle')
with open('df2.pickle','wb') as f:
    pickle.dump(df, f)

# Load
df1 = pd.read_pickle('df1.pickle')
with open('df2.pickle','rb') as f:
    df2 = pickle.load(f)

# Is ok?
if (df1 == df2).all().all():
    print('Data is ok.')

# Load opposite
df3 = pd.read_pickle('df2.pickle')
with open('df1.pickle','rb') as f:
    df4 = pickle.load(f)

# Is ok?
if (df3 == df4).all().all():
    print('Data opposite is ok.')

结果：

Data is ok.
Data opposite is ok.

有什么不同吗？我看到一些不同的输出pickle文件大小。Pandas版本的文件更大。

-rw-rw-r-- 1 spasz spasz 694 lis 15 17:38 df1.pickle
-rw-rw-r-- 1 spasz spasz 662 lis 15 17:38 df2.pickle

在python 3.8.0和panda 1.5.0上进行了测试。