Adding a new custom column in a python data frame [closed]

Question

I need to add a new column in a python data frame that has the dates of January in a column. Each date is repeated 24 times in the column without any intervention. The total number of entries will thus be 31*24 = 744 entries. Can you please help me code this part ?

Answer 1

This turns out to be fairly simple! There is a handy method called repeat on a datetime index. Here are the steps:

import pandas as pd

Define a date range, supplying start and end

jan = pd.date_range(start="1-Jan-2018", end="31-Jan-2018")    # could specify any year

Now provide how many times you want to repeat each date and create the repeated column

num_repeats = 24
repeated_jans = jan.repeat(num_repeats)

Let's create the random dummy dataframe as a base

total_dates = num_repeats * len(jan)    # 24 x 31 = 744
df = pd.DataFrame(np.random.randint(0, 10, total_dates))

This is how we add a column - the name of the column can be anything

df['repeated_jans_lalala'] = repeated_jans

Have a look at some dates:

print(df.iloc[[0, 24, 48, 71, 72]])    # multiples of 24...we can see one repeated date

    0 repeated_jans_lalala
0   7    2018-01-01
24  4    2018-01-02
48  3    2018-01-03
71  6    2018-01-03
72  3    2018-01-04

If try adding the column to a dataframe with a different number of rows, we get an error:

df_error = pd.DataFrame(np.random.randint(0, 10, 743))    # require 744!
#df_error['repeated_jans'] = repeated_jans                         # raises ValueError

If you want to change the way the dates look, you can use the strftime method on the dates

jans_fancy = jans.strftime('%d-%B-%y')
df['fancy_jans'] = jans_fancy

df.head()

0   repeated_jans_lalal  fancy_jans
0   7   2018-01-01       01-January-18
1   2   2018-01-01       01-January-18
2   8   2018-01-01       01-January-18
3   9   2018-01-01       01-January-18
4   7   2018-01-01       01-January-18

If you don't want to show the actual year, just leave out the %y part!