使用pandas中的round和compare

import pandas as pd

d1 = {"col": [7.1, 2.0, 3.0, 4.0, None, 1.9, 1.3]}
d2 = {"col": [7.1, 2.5, 3.0, 4.0, None, 1.2, None]}

df1 = pd.DataFrame(d1)
df2 = pd.DataFrame(d2)

# to deal with the precision issue, say we want to round to 4 decimal place
df1["col"] = df1["col"].round(4)
df2["col"] = df2["col"].round(4)

# returns a dataframe with the differences
comparison = df1.compare(df2)

print(comparison)
print("different rows count:", len(comparison))

字符串
输出将是

col      
  self other
1  2.0   2.5
5  1.9   1.2
6  1.3   NaN
different rows count: 3

型

使用numpy.isclose和equal_nan=True参数，如果需要，通过rtol参数设置精度，最后计数True s使用sum：

df = pd.DataFrame({'a':[np.nan, 10.12],
                   'b':[np.nan, 10.121231]})

out = np.isclose(df.a, df.b, equal_nan=True, rtol = 0.003)
print (out)
[ True  True]

print (out.sum())
2

different = len(out) - out.sum()
0

字符串