将带字母的时间戳转换为日期时间

Question

I have a txt file with data and values like this one:

PP  C   timestamp   HR  RMSSD   SCL 
PP1 1   20120918T131600000  NaN NaN 80.239727 
PP1 1   20120918T131700000  61  0.061420    77.365127

and I am importing it like that:

df = pd.read_csv('data.txt','\t', header=0)

which gives me a nice looking dataframe:

Running

df.columns

shows this result Index(['PP', 'C', 'timestamp', 'HR', 'RMSSD', 'SCL'], dtype='object') .

Now when I am trying to convert the timestamp column into a datetime column:

df["datetime"] = pd.to_datetime(df["timestamp"], format='%Y%m%dT%H%M%S%f')

I get this: ValueError: time data 'timestamp' does not match format '%Y%m%dT%H%M%S%f' (match)

Any ideas would be appreciated.

Answer 1

First, the error message you're quoting is from the header row. It's trying to parse the literal string 'timestamp' as a timestamp, which is failing. If you're getting an error on an actual data row, show us that message.

All three of your posted data rows parse fine with your format in my testing:

>>> [pandas.to_datetime(s, format='%Y%m%dT%H%M%S%f') 
    for s in ['20120918T131600000', '20120918T131700000', 
              '20120918T131800000']]
[Timestamp('2012-09-18 13:16:00'), Timestamp('2012-09-18 13:17:00'), Timestamp('2012-09-18 13:18:00')]

I have no idea where you got format='%Y%m%dT%H%M%S%f'[:-3] , which just removes the S%f from the format string, leaving it invalid. If you want to remove the last three digits of the data so that you ca just use %H%M%S instead of %H%M%S%f , you need to put the [:-3] on the timestamp data value, not the format.