Converting dates from DMY format to DM format

Asaf Yancu

Join Date: May 2021

Posts: 38
#1

Converting dates from DMY format to DM format

22 Nov 2023, 10:31

Hi all Statalists!
I have dates in my data set in DMY format. Here is a subset of the data:

Code:

* Example generated by -dataex-. For more info, type help dataex clear input int test_date 16560 16579 16589 16589 16602 16615 16618 15465 15467 15467 end format %td test_date

I want to generate a new variable that takes the day and the month of the corresponding full date (test_date).
For example, if test_date == 16560 (04may2005 in DMY format), the new variable takes 04may (in DM format).
I need this new variable for regression analysis.

Many thanks!
Tags: None
Nick Cox

Join Date: Mar 2014

Posts: 35698
#2

22 Nov 2023, 10:52

You can always specify a display format that suppresses the year.

Code:

. di %tddd_Mon mdy(11, 2, 2023) 2 Nov . di %tddd_Mon mdy(11, 22, 2023) 22 Nov .

Otherwise, don't confuse what is stored with what is displayed. See e.g. https://journals.sagepub.com/doi/pdf...867X1201200415

If you want day of year, use the function doy().
Comment

Asaf Yancu

Join Date: May 2021
Posts: 38

23 Nov 2023, 02:24

Thanks Nick!
I think the function doy() is close to what I need, but not exactly.
Here is an exmple:

Code:

* Example generated by -dataex-. For more info, type help dataex
clear
input int test_date float doy
16560 124
16194 124
16560 124
16560 124
16560 124
16560 124
16560 124
16560 124
16560 124
16560 124
end
format %td test_date

I generated the variable "doy" using the doy function

Code:

gen doy = doy(test_date)

However, 04may2005 and 03may2004 get the same value of doy:

Code:

. list

     +-----------------+
     | test_date   doy |
     |-----------------|
  1. | 04may2005   124 |
  2. | 03may2004   124 |
  3. | 04may2005   124 |
  4. | 04may2005   124 |
  5. | 04may2005   124 |
     |-----------------|
  6. | 04may2005   124 |
  7. | 04may2005   124 |
  8. | 04may2005   124 |
  9. | 04may2005   124 |
 10. | 04may2005   124 |
     +-----------------+

My dataset includes exam dates from 2000 to 2005. Thus, for instance, I need to assign the same value for all exams that were taken on May 4th (May 4th, 2000, May 4th, 2001, ..., May 4th, 2005) and another for all exams that were taken on May 3rd.

Last edited by Asaf Yancu; 23 Nov 2023, 02:27.

Comment

Nick Cox

Join Date: Mar 2014
Posts: 35698

23 Nov 2023, 03:34

The reason you get different values for 2004 and 2005 is naturally that 2004 was a leap year and 2005 was not. You will get identical values for any particular day of year if you use the same year as reference. If you have any instances of 29th February you must use a leap year as reference. The price to be paid is a jump of 2 not 1 from February 28 to March 1 in any non-leap year.

For example

Code:

* Example generated by -dataex-. For more info, type help dataex
clear
input int test_date float doy
16560 124
16194 124
16560 124
16560 124
16560 124
16560 124
16560 124
16560 124
16560 124
16560 124
end
format %td test_date

gen wanted = doy(mdy(month(test_date), day(test_date), 2024)) 

list 

    +--------------------------+
     | test_date   doy   wanted |
     |--------------------------|
  1. | 04may2005   124      125 |
  2. | 03may2004   124      124 |
  3. | 04may2005   124      125 |
  4. | 04may2005   124      125 |
  5. | 04may2005   124      125 |
     |--------------------------|
  6. | 04may2005   124      125 |
  7. | 04may2005   124      125 |
  8. | 04may2005   124      125 |
  9. | 04may2005   124      125 |
 10. | 04may2005   124      125 |
     +--------------------------+

Announcement