CODEWITHSUNDEEP
pythonpandas
user3000538
user3000538

Reputation: 203

Converting pandas column to a user defined week numbers

Below is a small subset of a quite big dataframe in python.

ID          invoiceDate
1           2017-07-18
2           2017-10-17
3           2017-08-20
4           2017-09-05
5           2017-08-20
6           2017-11-08
7           2017-09-17
8           2017-11-10
9           2017-08-14
10          2017-08-24

Given that the period of interest is between 2017-07-13 and 2017-11-21, it is equivalent to 20 weeks. I need to map each value in invoiceDate to the corresponding week number. For example, 2017-07-18 is mapped to week #1 (week one starts and ends at 2017-07-13 and 2017-07-19, respectively) or 2017-08-20 to week #6 (week six starts and ends at 2017-08-17 and 2017-08-23, respectively). The invoiceDate is of type object.

Upvotes: 1

Views: 535

Answers (3)

Ankur Sinha
Ankur Sinha

Reputation: 6639

Are you looking for something like this?

df = pd.DataFrame()
df['ID'] = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
df['InvDate'] = ['2017-07-18', '2017-10-17', '2017-08-20', '2017-09-05', '2017-08-20', '2017-11-08', '2017-09-17', '2017-11-10', '2017-08-14', '2017-08-24']

df['InvDate'] = pd.to_datetime(df['InvDate'])
df['WeekNum'] = df['InvDate'].dt.to_period('W').dt.week
df['WeekNum'] = df['WeekNum'] - df['WeekNum'][0] + 1

Output:

   ID    InvDate  WeekNum
0   1 2017-07-18        1
1   2 2017-10-17       14
2   3 2017-08-20        6
3   4 2017-09-05        8
4   5 2017-08-20        6
5   6 2017-11-08       17
6   7 2017-09-17       10
7   8 2017-11-10       18
8   9 2017-08-14        5
9  10 2017-08-24        7

Upvotes: 1

Scott Boston
Scott Boston

Reputation: 153460

Use to_period with freq 'W-Wed' altering the start of the week.

s = df['invoiceDate'].dt.to_period('W-Wed').dt.week
df['Week_Num'] = (s - s.min() + 1)

df['Week_Num'] = (df['invoiceDate'].dt.to_period('W-Wed').dt.week - df['invoiceDate'].dt.to_period('W-Wed').dt.week.min() + 1)

Output:

   productID invoiceDate  Week_Num
0          1  2017-07-18         1
1          2  2017-10-17        14
2          3  2017-08-20         6
3          4  2017-09-05         8
4          5  2017-08-20         6
5          6  2017-11-08        17
6          7  2017-09-17        10
7          8  2017-11-10        18
8          9  2017-08-14         5
9         10  2017-08-24         7

Upvotes: 5

BENY
BENY

Reputation: 323226

IIUC

(df.invoiceDate-pd.to_datetime('2017-07-13')).dt.days//7+1
Out[352]: 
0     1
1    14
2     6
3     8
4     6
5    17
6    10
7    18
8     5
9     7
Name: invoiceDate, dtype: int64

Upvotes: 2

Related Questions