Add a date column in pandas df using constant value in str

Question:

i have a table in pandas df

    product_id_x    product_id_y    count
0   2727846            7872456       1
1   29234              2932348       2
2   29346              9137500       1
3   29453              91365738      1
4   2933666            91323494      1

i want to add a new column ‘dates’ which i have defined in a str.

dateSelect = "'2016-11-06'"

so i added a new constant column

df['dates'] = dateSelect

but i am getting result as 

   product_id_x   product_id_y    count   dates
0   2727846          7872456         1  '2016-11-06'
1   29234            2932348         2  '2016-11-06'
2   29346            9137500         1  '2016-11-06'
3   29453            91365738        1  '2016-11-06'
4   2933666          91323494        1  '2016-11-06'

the values in the dates are coming in quotes. and 

type(df['dates']) = str

but i want it in date format, because further i am going to store this table in my mysql database. and i want the type to be date.

from sqlalchemy import create_engine
engine = create_engine('mysql+mysqldb://name:[email protected]/dbname', echo=False)
df.to_sql(name='tablename', con=engine, if_exists = 'append', index=False)
Asked By: Shubham R

||
Source

Answers:

I think you can use first replace ' by empty space and then to_datetime:

dateSelect = pd.to_datetime("'2016-11-06'".replace("'",""))
print (dateSelect)
2016-11-06 00:00:00

print (type(dateSelect))
<class 'pandas.tslib.Timestamp'>

df['dates'] = pd.to_datetime("'2016-11-06'".replace("'",""))

print (df)
   product_id_x  product_id_y  count      dates
0       2727846       7872456      1 2016-11-06
1         29234       2932348      2 2016-11-06
2         29346       9137500      1 2016-11-06
3         29453      91365738      1 2016-11-06
4       2933666      91323494      1 2016-11-06

print (df.dtypes)
product_id_x             int64
product_id_y             int64
count                    int64
dates           datetime64[ns]
dtype: object
Answered By: jezrael

In it don’t put double quote avoiding to define it as string. 

dateSelect = '2016-11-06'  
df['dates'] = dateSelect
Answered By: Chandan

Ahh! @jezrael got there first…

 print timeit.timeit("""
import pandas as pd
import datetime as dt
import timeit
df = pd.read_csv('date_time_pandas.csv')
dateSelect_str = "2016-11-06"

# using standard datetime
dateSelect = dt.datetime.strptime(dateSelect_str,"%Y-%m-%d")
df['dates'] = dateSelect
#print(df['dates'])
""",number=100)


# Alternate method using pandas datetime
print timeit.timeit("""
import pandas as pd
import datetime as dt
import timeit
df = pd.read_csv('date_time_pandas.csv')
dateSelect_str = "2016-11-06"

dateSelect = pd.to_datetime(dateSelect_str, format='%Y-%m-%d', errors='ignore')
df['dates'] = dateSelect
#print df['dates']
""",number=100)

gives output – 

0.228258825751
0.167258402887

on an average.

Conclusion Using pd_datetime in this case is more efficient

Answered By: Vivek Kalyanarangan

most direct route

df['dates'] = pd.Timestamp('2016-11-06')
df

   product_id_x  product_id_y  count      dates
0       2727846       7872456      1 2016-11-06
1         29234       2932348      2 2016-11-06
2         29346       9137500      1 2016-11-06
3         29453      91365738      1 2016-11-06
4       2933666      91323494      1 2016-11-06
Answered By: piRSquared

For date range pd.date_range function is the best one..

dataFrame["Date Column"] = pd.date_range("1/08/2020", periods=len(dataFrame))

All the best!

Answered By: Aamer