Conversion of Strings to Datetime Pandas - tried everything

Question

I know there are a lot of similar questions but I am pretty sure I have tried everything I found

Data:

    TimeStamp
    10:34:30
    10:35:30
    10:36:30
    10:37:30

I have two different .csv files and I am trying to compare the times from the two in order determine further action.

I want to use .dt.hour or .dt.min to compare the times. Here's what I tried: The type always remains str or float and the .dt.hour or .dt.min does not work

    def csv_to_df(path, filename):
        dateparse = lambda x: pd.datetime.strptime(x, '%H:%M:%S')
        df = pd.read_csv(path + filename, skipinitialspace = True, parse_dates = True, date_parser = dateparse)
        return df

I also tried using pd.to_datetime later for conversion. That didn't work either.

    df_sr4500.TimeStamp=pd.to_datetime(df_sr4500.TimeStamp).dt.time
    print(df_sr4500.TimeStamp[1].dt.hour)

    ERROR: AttributeError: 'datetime.time' object has no attribute 'dt'

@MattR yes in the form: df['Mycol'] = pd.to_datetime(df['Mycol'], format='%\H:%M:%S') — Brain_overflowed, Sep 29 '17 at 15:47
and what was that result? That may help us find out the issue. — MattR, Sep 29 '17 at 15:49
it gives me errors when I try to use df.iloc[0,0].dt.hour; it tells me that type 'str' doesn't have an attribute '.dt' — Brain_overflowed, Sep 29 '17 at 15:54
Please post that in the original question for others to see. It's hard to understand comment code — MattR, Sep 29 '17 at 15:55
Possible duplicate of [Convert Pandas Series to DateTime in a DataFrame](https://stackoverflow.com/questions/28133018/convert-pandas-series-to-datetime-in-a-dataframe) — MattR, Sep 29 '17 at 16:05

BENY · Accepted Answer · 2017-09-29T16:04:15.097

1

IIUC

df.TimeStamp=pd.to_datetime(df.TimeStamp).dt.time

df.TimeStamp[1]
Out[115]: datetime.time(10, 35, 30)

df.TimeStamp[1].hour

Out[119]: 10

In case you need convert new column hour

df.TimeStamp.apply(lambda x : x.hour)
Out[122]: 
0    10
1    10
2    10
3    10
Name: TimeStamp, dtype: int64

edited Sep 29 '17 at 16:04

answered Sep 29 '17 at 15:47

BENY

317,841
20
164
234

AttributeError: 'datetime.time' object has no attribute 'dt' – Brain_overflowed Sep 29 '17 at 15:52
Is your original data type ==string ? – BENY Sep 29 '17 at 15:54
@Brain_overflowed that is working on my side , and use your original df to test it – BENY Sep 29 '17 at 15:55
@Brain_overflowed after you figure out how to solve your problem , I will delete the answer. – BENY Sep 29 '17 at 16:02
@Brain_overflowed yw – BENY Sep 29 '17 at 16:03

score 0 · Answer 2 · answered Sep 29 '17 at 16:03

df.TimeStamp = pd.to_datetime(df_sr4500.TimeStamp) is sufficient.

To get the hour or the minute from this column you need the following:

[i.hour for i in df.TimeStamp]

Note: .dt access an individual item but you are trying to access it using an entire series.

Conversion of Strings to Datetime Pandas - tried everything

2 Answers2