How to destring date variable formatted as year and week number

[Yonatan Mesfin]

Dear Statalist,

I was just wondering if there is an easy way to destring a date variable formatted as "year and week number".
for example,

2008w34
2008w35
2008w36
2008w37
2008w38


Thanks so much for your help!

Yonatan

[Nick Cox]

You shouldn't want to destring that. How to do it follows from reading the help -- ignore("w") -- but that would mean values like 200834, which are not closer to Stata numeric date variables.

For more on why that is a bad idea, see

SJ-18-4 dm0098 . . Speaking Stata: Seven steps for vexatious string variables
. . . . . . . . . . . . . . . . . . . . N. J. Cox and C. B. Schechter
Q4/18 SJ 18(4):981--994 (no commands)
provides a step-by-step guide explaining how to convert string
variables or -- as the case may merit -- to leave them as they
are

A superficial answer would be that the function weekly() is what you want, as in

. clear

. input str7 myweek

        myweek
  1. 2008w34
  2. end

 gen wanted = weekly(myweek, "YW")

. format %tw wanted

. list

     +-------------------+
     |  myweek    wanted |
     |-------------------|
  1. | 2008w34   2008w34 |
     +-------------------+

But hang on! How are your weeks defined? In the same way as Stata weeks? I doubt it. One tell-tale sign is if that you have any weekly dates that are week 53.

For much more on weekly dates in Stata, see (for example) the recently active thread https://www.statalist.org/forums/for...-week-variable and its references.

For anything other than Stata weeks:

1. Work with the days of the week that define them.

2. Use delta(7) with tsset or xtset.

[Yonatan Mesfin]

Thanks Nick!

[Yonatan Mesfin]

Hi Nick,
May I ask you one more question?

I just wondering if it is possible to estimate exact end date of each week (Stata week) given that that specific week number is known or some other exact date in that given week is known.

Example;

Let say, one patient received flu vaccination on 03/01/2017. It is week 1. Is it possible to estimate the exact end date of week l based on the provided information in Stata?

I greatly appreciated your support.

Yonatan

[Nick Cox]

You could test various possibilities. I'd look first at whether you ever have week 0 and/or week 53 in your data.

It's better to drill down and get metadata from your data provider.