如何在pandas数据帧中使用if-else

Question

我已经开始学习大熊猫了,偶然发现了以下问题:

以下是一个包​​含以下数据的表:

书:

B_IDX B_NAME B_AUTHOR B_PRICE B_UTYPE B_ID
1     ABC    aaa      12.21   SCI     182
2     BCD    bbb      98      ECN     920
3     CDE    ccc      22.34   SCI     228
4     DEF    ddd      44.11   LIT     761
5     EFG    eee      0.99    MAT     10242
6     FGH    fff      4.99    MAT     77721

UCODE:

U_ID  U_CD
182   9982825
950   9992822
228   9999983
776   9912876
332   9003931

要求是使用if..else逻辑从上面提到的表中提取数据.

请求:

if B_UTYPE == 'SCI':
     pull the record from 'UCODE'
elif  B_UTYPE == 'MAT':
     split the  B_ID in 4 and 1 digits i.e. B_UTYPE.split[:2] and B_UTYPE.split[3:5]
else:
    keep the data as it is.

O/P例外:

B_ID B_NAME B_AUTHOR B_PRICE B_UTYPE B_ID   U_ID    U_CD      N_COL1  N_COL2
1    ABC    aaa      12.21   SCI     182    182     9982825   NA      NA
2    BCD    bbb      98      ECN     920    NA      NA        NA      NA
3    CDE    ccc      22.34   SCI     228    228     9999983   NA      NA
4    DEF    ddd      44.11   LIT     761    NA      NA        NA      NA 
5    EFG    eee      0.99    MAT     10242  NA      NA        102     42
6    FGH    fff      4.99    MAT     77721  NA      NA        777     21 

任何帮助/教程,我可以通过满足上述条件获得一些见解以达到预期的输出？

Answer 1

为了便于阅读,请分别构建每个结果,然后将各个部分连接在一起.

u_id = df.B_ID.astype(str).where(df.B_UTYPE.eq('SCI'))
u_cd = df.B_ID.map(ucode.set_index('U_ID').U_CD.astype(str))
ncol = (df.B_ID.astype(str)
          .str.extract(r'(\d{3})(\d+)')
          .where(df.B_UTYPE.eq('MAT'))
          .rename(columns=lambda x: f'N_COL{x+1}'))

df = pd.concat([df, u_id, u_cd, ncol], axis=1)
print(df)

   B_IDX B_NAME B_AUTHOR  B_PRICE B_UTYPE   B_ID B_ID     B_ID N_COL1 N_COL2
0      1    ABC      aaa    12.21     SCI    182  182  9982825    NaN    NaN
1      2    BCD      bbb    98.00     ECN    920  NaN      NaN    NaN    NaN
2      3    CDE      ccc    22.34     SCI    228  228  9999983    NaN    NaN
3      4    DEF      ddd    44.11     LIT    761  NaN      NaN    NaN    NaN
4      5    EFG      eee     0.99     MAT  10242  NaN      NaN    102     42
5      6    FGH      fff     4.99     MAT  77721  NaN      NaN    777     21

Answer 2

这是一个两步的方法.首先,您需要确定哪些行匹配哪个条件.然后,一旦有了条件和输出,就可以使用蒙版并将assign系列添加到DataFrame中.

c1 = book.B_UTYPE.eq("SCI")
c2 = book.B_UTYPE.eq("MAT")

s1 = book.B_ID.map(ucode.set_index('U_ID').U_CD)
s2 = book.B_ID.astype(str)

现在是有趣的部分:

parts = {
    'U_ID': book.B_ID.mask(~c1),
    'U_CD': pd.Series(s1).mask(~c1),
    'N_COL1': s2.str[:3].mask(~c2),
    'N_COL2': s2.str[3:].mask(~c2)      
}

book.assign(**parts)

   ID B_NAME B_AUTHOR  B_PRICE B_UTYPE   B_ID   U_ID       U_CD N_COL1 N_COL2
0   1    ABC      aaa    12.21     SCI    182  182.0  9982825.0    NaN    NaN
1   2    BCD      bbb    98.00     ECN    920    NaN        NaN    NaN    NaN
2   3    CDE      ccc    22.34     SCI    228  228.0  9999983.0    NaN    NaN
3   4    DEF      ddd    44.11     LIT    761    NaN        NaN    NaN    NaN
4   5    EFG      eee     0.99     MAT  10242    NaN        NaN    102     42
5   6    FGH      fff     4.99     MAT  77721    NaN        NaN    777     21

设置,以便您可以重现:

book = pd.DataFrame({'ID': {0: 1, 1: 2, 2: 3, 3: 4, 4: 5, 5: 6},
 'B_NAME': {0: 'ABC', 1: 'BCD', 2: 'CDE', 3: 'DEF', 4: 'EFG', 5: 'FGH'},
 'B_AUTHOR': {0: 'aaa', 1: 'bbb', 2: 'ccc', 3: 'ddd', 4: 'eee', 5: 'fff'},
 'B_PRICE': {0: 12.21, 1: 98.0, 2: 22.34, 3: 44.11, 4: 0.99, 5: 4.99},
 'B_UTYPE': {0: 'SCI', 1: 'ECN', 2: 'SCI', 3: 'LIT', 4: 'MAT', 5: 'MAT'},
 'B_ID': {0: 182, 1: 920, 2: 228, 3: 761, 4: 10242, 5: 77721}})

ucode = pd.DataFrame({'U_ID': {0: 182, 1: 950, 2: 228, 3: 776, 4: 332},
 'U_CD': {0: 9982825, 1: 9992822, 2: 9999983, 3: 9912876, 4: 9003931}})