Pandas-仅在根据特定条件检索的行上使用正则表达式提取带有数字/字符串的更新列

您可以extract使用IIUC与命名组，然后update：

df = pd.DataFrame({"Unit":["gm", np.NaN, np.NaN],
                   "Price":["40","80 per piece", "110 per pack"]})

  Unit         Price
0   gm            40
1  NaN  80 per piece
2  NaN  110 per pack

s = df.loc[df["Unit"].isnull(),"Price"].str.extract("(?P<Price>\d+)\sper\s(?P<Unit>[A-Za-z]+)").dropna()

df.update(s)

print (df)

    Unit Price
0     gm    40
1  piece    80
2   pack   110