I'm trying to convert a datetime column back to a string in Pandas dataframe.
the syntax I have so far is:
all_data['Order Day new'] = dt.date.strftime(all_data['Order Day new'], '%d/%m/%Y')
but this returns the error:
descriptor 'strftime' requires a 'datetime.date' object but received a 'Series'.
Can anyone tell me where I'm going wrong.
If you're using version 0.17.0 or higher then you can call this using .dt.strftime which is vectorised:
all_data['Order Day new'] = all_data['Order Day new'].dt.strftime('%Y-%m-%d')
** If your pandas version is older than 0.17.0 then you have to call apply and pass the data to strftime:
In [111]:
all_data = pd.DataFrame({'Order Day new':[dt.datetime(2014,5,9), dt.datetime(2012,6,19)]})
print(all_data)
all_data.info()
Order Day new
0 2014-05-09
1 2012-06-19
<class 'pandas.core.frame.DataFrame'>
Int64Index: 2 entries, 0 to 1
Data columns (total 1 columns):
Order Day new 2 non-null datetime64[ns]
dtypes: datetime64[ns](1)
memory usage: 32.0 bytes
In [108]:
all_data['Order Day new'] = all_data['Order Day new'].apply(lambda x: dt.datetime.strftime(x, '%Y-%m-%d'))
all_data
Out[108]:
Order Day new
0 2014-05-09
1 2012-06-19
In [109]:
all_data.info()
<class 'pandas.core.frame.DataFrame'>
Int64Index: 2 entries, 0 to 1
Data columns (total 1 columns):
Order Day new 2 non-null object
dtypes: object(1)
memory usage: 32.0+ bytes
You can't call strftime on the column as it doesn't understand Series as a param hence the error
1.0.5 this is producing a SettingWithCopyWarning.all_data['Order Day new']=all_data['Order Day new'].astype(str)
I think this is more simple, if the date is already in the format you want it in string form.
For converting all datetime columns to string use the below code.
for x in df.select_dtypes(include=['datetime64']).columns.tolist():
df[x] = df[x].astype(str)
or
date_columns = df.select_dtypes(include=['datetime64']).columns.tolist()
df[date_columns] = df[date_columns].astype(str)
In my case I had a pySpark dataframe. This is how I converted "day" of type timestamp, to a string:
import pyspark.sql.functions as F
convertedDf = rawDf.withColumn('d2', F.date_format(rawDf['day'], 'yyyyMMdd'))
