在DatetimeIndex上使用pandas时间戳对DataFrame进行切片

albert 发表于 Dev

阿尔伯特

我正在从磁盘将csv文件读取到pandas DataFrame中，并希望根据索引时间戳对数据帧进行切片/过滤。

到目前为止，这是我得到的：

INDEX_COL_NAME = 'Zeit'
DELIM_SIGN = ';'
DECIMAL_SIGN = ','
KEEP_COLUMNS = [-2]
ENCODING = 'ISO-8859-1'

DATE = (2016, 8, 11)
START = (10, 52, 0)
END = (10, 53, 0)

df = pd.read_csv('data.csv',
    delimiter=DELIM_SIGN,
    decimal=DECIMAL_SIGN,
    index_col=False,
    parse_dates=[INDEX_COL_NAME],
    infer_datetime_format=True,
    encoding=ENCODING)

df.set_index(INDEX_COL_NAME, inplace=True)
df = df[KEEP_COLUMNS]

date = pd.datetime(*DATE)
start = date.replace(hour=START[0], minute=START[1], second=START[2])
end = date.replace(hour=END[0], minute=END[1], second=END[2])

数据如下（缩短的代码段）：

Zeit;FU_P1;FU_P2;DIR_01;FIR_01;WAAGE_B1.I;WAAGE_B1.T;WAAGE_B1.X;WAAGE_B2.I;WAAGE_B2.T;WAAGE_B2.X;WAAGE_B3.I;WAAGE_B3.T;WAAGE_B3.X;WAAGE_B4.I;WAAGE_B4.T;WAAGE_B4.X;LEITFÄHIGKEIT_1.COND;LEITFÄHIGKEIT_2.COND
11.08.2016 10:51:59; 20,0; 0,00; 991,19;29,21;  0,0;Empty;   239;  1,0;Empty;-11,600;  0,0;Empty;-0,023;  0,0;Empty;-1,776;  0,3;Empty;
11.08.2016 10:52:00; 20,0; 0,00; 991,22;29,11;  0,0;Empty;   239;  1,0;Empty;-11,600;  0,0;Empty;-0,023;  0,0;Empty;-1,787;  0,3;Empty;
11.08.2016 10:52:10; 20,0; 0,00; 991,08;29,24;  0,0;Empty;   239;  1,0;Empty;-11,600;  0,0;Empty;-0,023;  1,0;Empty;-1,840;  0,3;Empty;
11.08.2016 10:52:20; 20,0; 0,00; 990,95;28,95;  0,0;Empty;   239;  1,0;Empty;-11,600;  0,0;Empty;-0,023;  0,0;Empty;-1,947;  0,3;Empty;
11.08.2016 10:52:30; 20,0; 0,00; 990,94;28,96;  0,0;Empty;   238;  1,0;Empty;-11,600;  0,0;Empty;-0,022;  0,0;Empty;-2,059;  0,3;Empty;
11.08.2016 10:52:40; 20,0; 0,00; 990,82;28,91;  0,0;Empty;   238;  1,0;Empty;-11,600;  0,0;Empty;-0,021;  0,0;Empty;-2,155;  0,3;Empty;
11.08.2016 10:52:50; 20,0; 0,00; 990,80;29,37;  0,0;Empty;   238;  1,0;Empty;-11,600;  0,0;Empty;-0,020;  0,0;Empty;-2,249;  0,0;Empty;
11.08.2016 10:53:00; 20,0; 0,00; 990,71;29,15;  0,0;Empty;   239;  1,0;Empty;-11,600;  0,0;Empty;-0,021;  1,0;Empty;-2,309;  0,5;Empty;
11.08.2016 10:53:01; 20,0; 0,00; 990,78;29,04;  0,0;Empty;   239;  1,0;Empty;-11,600;  0,0;Empty;-0,021;  0,2;Empty;-2,350;  0,5;Empty;

但是，由于以下原因，我无法获得所需的切片

print(df.ix[start:end]

打印一个空的DataFrame。

元素是索引的一部分，如

print(df.index)

表演

DatetimeIndex(['2016-11-08 10:45:27', '2016-11-08 10:45:28',
           '2016-11-08 10:45:29', '2016-11-08 10:45:30',
           '2016-11-08 10:45:31', '2016-11-08 10:45:32',
           '2016-11-08 10:45:33', '2016-11-08 10:45:34',
           '2016-11-08 10:45:35', '2016-11-08 10:45:36',
           ...
           '2016-11-08 15:59:51', '2016-11-08 15:59:52',
           '2016-11-08 15:59:53', '2016-11-08 15:59:54',
           '2016-11-08 15:59:55', '2016-11-08 15:59:56',
           '2016-11-08 15:59:57', '2016-11-08 15:59:58',
           '2016-11-08 15:59:59', '2016-11-08 16:00:00'],
          dtype='datetime64[ns]', name='Zeit', length=10408, freq=None)

每秒钟都有行，形成从数据记录开始到结束的总时间。

此外