python - 如何将单个 pandas.DataFrame 行拆分为多个由空格分隔的列？ Python

Question

我正在处理 NCEI 海洋数据，这些数据是没有标题的 .dat 文件，使用 python(https://www.ncei.noaa.gov/data/marine/icoads3.0/ 用于文件)它们看起来像:

166210151200 4962 35378 1306    101134     NL 1585 26                                                        165 17796730133 5 0                   2FF11FF11AAAAAAAAAAAA     98150000003002199 0 NAN     NATIONAAL ARCHIEF OF THE NETHERLANDS              DEN HAAG  NEDERLAND     1.11.01.01     1229             AANW                                   112                            AAN_1229_112                  DUTCH                         0493800N 102600E493700N         2 1TENERIFE                                                                                                                                                                                                                                                           0 21662101512         3         VM    8UNKNOWN        MAARSEVEEN                    DUTCH                  VOC                               M. GERRITSZ. BOOS             OPPERSTUURMAN                                                                                                                          ROTTERDAM                                    BATAVIA                                           0                                                  0977.216621015WZW                    7.00                                                                                                        UNKNOWN        UNKNOWN                                       UNKNOWN360 DEGREES                                                                                                                                                                                                 ZZO                                                                                                                                                                            MOU (?) KOELTE                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      00000000CLIWOC VERSION 1.0
166210161300 4907 35215 1306    101134     NL 1585 26                                                        165 17797730133 5 0                   2FF11FF11AAAAAAAAAAAA     98150000013002199 0 NAN     NATIONAAL ARCHIEF OF THE NETHERLANDS              DEN HAAG  NEDERLAND     1.11.01.01     1229             AANW                                   112                            AAN_1229_112                  DUTCH                         0490400N  84800E                1 1TENERIFE                                                                                                                                                                                                                                                           0 21662101612         3         VM    8UNKNOWN        MAARSEVEEN                    DUTCH                  VOC                               M. GERRITSZ. BOOS             OPPERSTUURMAN                                                                                                                          ROTTERDAM                                    BATAVIA                                           0                                                  0977.216621015ZW 1/2 N              18.00                                                                                                        UNKNOWN        UNKNOWN                                       UNKNOWN360 DEGREES                                                                                                                                                                                                 ZZO                                                                                                                                                                            MOU KOELTE                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          00000000CLIWOC VERSION 1.0
166210171300 4812 35000 1306    101134     NL 1695 26                                                        165 17680730133 5 0                   2FF11FF11AAAAAAAAAAAA     98150000023002199 0 NAN     NATIONAAL ARCHIEF OF THE NETHERLANDS              DEN HAAG  NEDERLAND     1.11.01.01     1229             AANW                                   112                            AAN_1229_112                  DUTCH                         0483000N  63900E480700N         2 1TENERIFE                                                                                                                                                                                                                                                           0 21662101712         3         VM    8UNKNOWN        MAARSEVEEN                    DUTCH                  VOC                               M. GERRITSZ. BOOS             OPPERSTUURMAN                                                                                                                          ROTTERDAM                                    BATAVIA                                           0                                                  0977.216621015ZWTW                  15.00                                                                                                        UNKNOWN        UNKNOWN                                       UNKNOWN360 DEGREES                                                                                                                                                                                                 ZTO                                                                                                                                                                            MOU KOELTE                                                                                                                                                                                                                                        MOOI WEER                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         00000000CLIWOC VERSION 1.0
166210181300 4758 34925 1306    101134     NL 1695 26                                                        165 17670730133 5 0                   2FF11FF11AAAAAAAAAAAA     98150000033002199 0 NAN     NATIONAAL ARCHIEF OF THE NETHERLANDS              DEN HAAG  NEDERLAND     1.11.01.01     1229             AANW                                   112                            AAN_1229_112                  DUTCH                         0474100N  55400E473500N         2 1TENERIFE                                                                                                                                                                                                                                                           0 21662101812         3         VM    8UNKNOWN        MAARSEVEEN                    DUTCH                  VOC                               M. GERRITSZ. BOOS             OPPERSTUURMAN                                                                                                                          ROTTERDAM                                    BATAVIA                                           0                                                  0977.216621015ZWTW                  11.00                                                                                                        UNKNOWN        UNKNOWN                                       UNKNOWN360 DEGREES                                                                                                                                                                                                 ZTO                                                                                                                                                                            MOU KOELTE                                                                                                                                                                                                                                        'ENN MOUT'?                                                                                                                                           REGEN                                                                                                                                                                                                                                                                                                                                                                                                       01000000CLIWOC VERSION 1.0
166210191300 4757 34795 1306    101134     NL 1805 67                                                        165 17672730133 5 0                   2FF11FF11AAAAAAAAAAAA     98150000043002199 0 NAN     NATIONAAL ARCHIEF OF THE NETHERLANDS              DEN HAAG  NEDERLAND     1.11.01.01     1229             AANW                                   112                            AAN_1229_112                  DUTCH                         0473400N  43600E                1 1TENERIFE                                                                                                                                                                                                                                                           0 21662101912         3         VM    8UNKNOWN        MAARSEVEEN                    DUTCH                  VOC                               M. GERRITSZ. BOOS             OPPERSTUURMAN                                                                                                                          ROTTERDAM                                    BATAVIA                                           0                                                  0977.216621015W/Z                   14.00                                                                                                        UNKNOWN        UNKNOWN                                       UNKNOWN360 DEGREES                                                                                                                                                                                                 Z                                                                                                                                                                              MARSZEILSKOELTE, TOUPKOULTE                                                                                                                                                                                                                                                                                                                                                                             REGEN                                                                                                                                                                                                                                                                                                                                                                                                       01000000CLIWOC VERSION 1.0

这些是我一直使用的制表符分隔文件

data = pd.read_table('file.dat', header=None)

它将数据导入为 x 行，其中包含所有数据的单个列。在单列中，每个数据由空格分隔。

有没有一种方法可以将这些数据导入列或读取数据变量并根据空白将每一行拆分为列。我认为这就是我对 read.table 函数所做的。完整的数据集很大，所以我更喜欢一种方法来导入它们，而不是之后必须处理它们。

Answer 1

我想你需要的是Fixed Width Formatted :

代码:

df = pd.read_fwf('IMMA.dat', header=None)
print(df.dtypes)

结果:

[17 rows x 66 columns]
0       int64
1       int64
2       int64
3       int64
       ...   
61     object
62     object
63     object
64     object
65    float64
dtype: object