熊猫用非数值减去两个数据帧

使用 merge indicator=True ,然后按 query 拆下立柱 _merge :

df1['c'] = df1['c'].apply(tuple)
df2['c'] = df2['c'].apply(tuple)

df3 = (df2.merge(df, how='left', indicator=True)
          .query('_merge == "left_only"')
          .drop('_merge', axis=1))

df3['c'] = df3['c'].apply(list)
print (df3)
    a    b         c   d
2  14  Dil  [11, 18]  ZZ

concat drop_duplicates

前任:

import pandas as pd

df = pd.DataFrame({"a": [12, 13], "b":["vik", "Rah"], "c":[[9,  18], [10, 18]], "d":["SS", "YY"]})
df2 = pd.DataFrame({"a": [12, 13, 14], "b":["vik", "Rah", "Dil"], "c":[[9,  18], [10, 18], [11, 18]], "d":["SS", "YY", "ZZ"]})

df3 = pd.concat([df, df2], ignore_index=True)
df3["c"] = df3["c"].apply(tuple)
print(df3.drop_duplicates(keep=False))