熊貓數據框根據其他列值的范圍插入值

Question

我有如下數據框，我想根據sic2列中的值插入一個“字符串”。

        conm            sic2
115466  ALLEGION PLC    34.0
115471  AGILITY HEALTH INC  80.0
115473  NORDIC AMERICAN OFFSHORE    44.0
115474  AAD             54.0
115477  DORIAN LPG LTD  44.0
115484  NOMAD FOODS LTD 20.0
115486  ATHENE HOLDING LTD  63.0
115490  MIDATECH PHARMA PLC 28.0
115495  MOTIF BIO PLC   28.0

sic2數字到字符串的范圍如下。

1-9 Agriculture, Forestry and Fishing
10-14   Mining
15-17   Construction
18-19   not used
20-39   Manufacturing
40-49   Transportation, Communications, Electric, Gas and Sanitary service
50-51   Wholesale Trade
52-59   Retail Trade
60-67   Finance, Insurance and Real Estate
70-89   Services
91-97   Public Administration
99-99   Nonclassifiable
0 -1    Agricultural Production-Crops

如何制作像這樣的pandas.DataFrame應用整個大型數據集？

我嘗試了幾種條件代碼，但是它一直失敗。

        conm            sic2                industry
115466  ALLEGION PLC    34.0                Manufacturing
115471  AGILITY HEALTH INC  80.0            Services
115473  NORDIC AMERICAN OFFSHORE    44.0    Transportation, Communications, Electric, Gas and Sanitary service
115474  AAD             54.0                Retail Trade

Answer 1

如果您將sics數字轉換成字典，那么根據需要查找行業是相當簡單的：

碼：

sic = [x.strip().split(' ', 1) for x in """
    1-9 Agriculture, Forestry and Fishing
    10-14 Mining
    15-17 Construction
    18-19 not used
    20-39 Manufacturing
    40-49 Transportation, Communications, ...
    50-51 Wholesale Trade
    52-59 Retail Trade
    60-67 Finance, Insurance and Real Estate
    70-89 Services
    91-97 Public Administration
    99-99 Nonclassifiable
""".split('\n')[1:-1]]

sic_dict = dict(sum([[(x, z) for x in
                      range(*[int(y) for y in v.split('-')])]
                     for v, z in sic], []))

測試代碼：

df = pd.read_fwf(StringIO(u"""
    number  conm                      sic2
    115466  ALLEGION PLC              34.0
    115471  AGILITY HEALTH INC        80.0
    115473  NORDIC AMERICAN OFFSHORE  44.0
    115474  AAD                       54.0
    115477  DORIAN LPG LTD            44.0
    115484  NOMAD FOODS LTD           20.0
    115486  ATHENE HOLDING LTD        63.0
    115490  MIDATECH PHARMA PLC       28.0
    115495  MOTIF BIO PLC             28.0"""), header=1)

df['industry'] = df.sic2.apply(lambda x: sic_dict[int(x)])

print(df)

結果：

   number                      conm  sic2                             industry
0  115466              ALLEGION PLC  34.0                        Manufacturing
1  115471        AGILITY HEALTH INC  80.0                             Services
2  115473  NORDIC AMERICAN OFFSHORE  44.0  Transportation, Communications, ...
3  115474                       AAD  54.0                         Retail Trade
4  115477            DORIAN LPG LTD  44.0  Transportation, Communications, ...
5  115484           NOMAD FOODS LTD  20.0                        Manufacturing
6  115486        ATHENE HOLDING LTD  63.0   Finance, Insurance and Real Estate
7  115490       MIDATECH PHARMA PLC  28.0                        Manufacturing
8  115495             MOTIF BIO PLC  28.0                        Manufacturing

Answer 2

#Save your mapping table to a data frame

df2 = pd.DataFrame({'id_end': {0: 9,  1: 14,  2: 17,  3: 19,  4: 39,  5: 49,  6: 51,  7: 59,  8: 67,  9: 89,  10: 97,  11: 99,  12: 1},
 'id_start': {0: 1,  1: 10,  2: 15,  3: 18,  4: 20,  5: 40,  6: 50,  7: 52,  8: 60,  9: 70,  10: 91,  11: 99,  12: 0},
 'industry': {0: 'Agriculture, Forestry and Fishing',  1: 'Mining',  2: 'Construction',  3: 'not used',  4: 'Manufacturing',
  5: 'Transportation, Communications, Electric, Gas and Sanitary service',
  6: 'Wholesale Trade',  7: 'Retail Trade',  8: 'Finance, Insurance and Real Estate',  9: 'Services',  
  10: 'Public Administration',  11: 'Nonclassifiable',  12: 'Agricultural Production Crops'}})

df2 = df2.sort_values(by='id_end')

Out[354]: 
    id_end  id_start                                           industry
12       1         0                      Agricultural Production Crops
0        9         1                  Agriculture, Forestry and Fishing
1       14        10                                             Mining
2       17        15                                       Construction
3       19        18                                           not used
4       39        20                                      Manufacturing
5       49        40  Transportation, Communications, Electric, Gas ...
6       51        50                                    Wholesale Trade
7       59        52                                       Retail Trade
8       67        60                 Finance, Insurance and Real Estate
9       89        70                                           Services
10      97        91                              Public Administration
11      99        99                                    Nonclassifiable

#Map sic2 number to industry names
df['industry'] = df['sic2'].astype(np.int).apply(lambda x: df2.loc[df2.id_end>=x,'industry'].iloc[0])


Out[352]: 
                            conm  sic2                                             industry
115466              ALLEGION PLC  34.0                                        Manufacturing 
115471        AGILITY HEALTH INC  80.0                                             Services 
115473  NORDIC AMERICAN OFFSHORE  44.0    Transportation, Communications, Electric, Gas ... 
115474                       AAD  54.0                                         Retail Trade 
115477            DORIAN LPG LTD  44.0    Transportation, Communications, Electric, Gas ... 
115484           NOMAD FOODS LTD  20.0                                        Manufacturing 
115486        ATHENE HOLDING LTD  63.0                   Finance, Insurance and Real Estate 
115490       MIDATECH PHARMA PLC  28.0                                        Manufacturing 
115495             MOTIF BIO PLC  28.0                                        Manufacturing