How do i convert a pandas index of strings to datetime format
my dataframe 'df' is like this
value
2015-09-25 00:46 71.925000
2015-09-25 00:47 71.625000
2015-09-25 00:48 71.333333
2015-09-25 00:49 64.571429
2015-09-25 00:50 72.285714
but the index is of type string, but i need it a datetime format because i get the error
'Index' object has no attribute 'hour'
when using
df['A'] = df.index.hour
The answer is
It should work as expected. Try to run the following example.
import pandas as pd
import io
data = """value
"2015-09-25 00:46" 71.925000
"2015-09-25 00:47" 71.625000
"2015-09-25 00:48" 71.333333
"2015-09-25 00:49" 64.571429
"2015-09-25 00:50" 72.285714"""
df = pd.read_table(io.StringIO(data), delim_whitespace=True)
# Converting the index as date
df.index = pd.to_datetime(df.index)
# Extracting hour & minute
df['A'] = df.index.hour
df['B'] = df.index.minute
df
# value A B
# 2015-09-25 00:46:00 71.925000 0 46
# 2015-09-25 00:47:00 71.625000 0 47
# 2015-09-25 00:48:00 71.333333 0 48
# 2015-09-25 00:49:00 64.571429 0 49
# 2015-09-25 00:50:00 72.285714 0 50
In my case, my dataframe has the following characteristics
<class 'pandas.core.frame.DataFrame'>
Index: 3040 entries, 15/12/2008 to
Data columns (total 1 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Close 3038 non-null float64
dtypes: float64(1)
memory usage: 47.5+ KB
The first option data.index = pd.to_datetime(data.index) returned
ParserError: String does not contain a date: ParserError: String does not contain a date:
The second option: data.index.to_datetime() returned
AttributeError: 'Index' object has no attribute 'to_datetime'
It returned
Another option I have tested is. data.index = pd.to_datetime(data.index)
It returned: ParserError: String does not contain a date:
What could be my problem? Thanks
I just give other option for this question - you need to use '.dt' in your code:
import pandas as pd_x000D_
_x000D_
df.index = pd.to_datetime(df.index)_x000D_
_x000D_
#for get year_x000D_
df.index.dt.year_x000D_
_x000D_
#for get month_x000D_
df.index.dt.month_x000D_
_x000D_
#for get day_x000D_
df.index.dt.day_x000D_
_x000D_
#for get hour_x000D_
df.index.dt.hour_x000D_
_x000D_
#for get minute_x000D_
df.index.dt.minuteYou could explicitly create a DatetimeIndex when initializing the dataframe. Assuming your data is in string format
data = [
('2015-09-25 00:46', '71.925000'),
('2015-09-25 00:47', '71.625000'),
('2015-09-25 00:48', '71.333333'),
('2015-09-25 00:49', '64.571429'),
('2015-09-25 00:50', '72.285714'),
]
index, values = zip(*data)
frame = pd.DataFrame({
'values': values
}, index=pd.DatetimeIndex(index))
print(frame.index.minute)
Doing
df.index = pd.to_datetime(df.index, errors='coerce')
the data type of the index has changed to
Source: Stackoverflow.com