Is there a way to format variable?

[Yuchen Li]

Hello friends,

I am currently using data that has one variable that shows the value of the date, for example, a variable named effective_date shows the value 20120105 means 2012-January-5th. If I only need the value of year (for my analysis purpose, as I need to create an arrival cohort based on years, not the exact date).

Right now I recode it like (20120000/20129999 = 2012), It works, but I am curious is there a way to formate variable just show the first four digitals instead of recoding it?

Please advise

Thanks

[William Lisowski]

Formatting the date variable to only present the first four digits will not solve your analytic problem, because Stata doesn't pay attention to the format and will treate 20120106 as different than 20120107.

The easy way to get the year from your numeric variable is to use mathematics.

generate year = floor(date/10000)

. * Example generated by -dataex-. For more info, type help dataex
. clear

. input long date

             date
  1. 20120105
  2. 20211231
  3. end

. generate year = floor(date/10000)

. list, clean

           date   year  
  1.   20120105   2012  
  2.   20211231   2021  

.

For documentation

help floor()

[Yuchen Li]

Formatting the date variable to only present the first four digits will not solve your analytic problem, because Stata doesn't pay attention to the format and will treate 20120106 as different than 20120107.

The easy way to get the year from your numeric variable is to use mathematics.

generate year = floor(date/10000)

. * Example generated by -dataex-. For more info, type help dataex
. clear

. input long date

date
1. 20120105
2. 20211231
3. end

. generate year = floor(date/10000)

. list, clean

date year
1. 20120105 2012
2. 20211231 2021

.

For documentation

help floor()

Thank you so much for your help William!

[Nick Cox]

William Lisowski gave a fine solution. Note that date functions will work here too given a twist.

. clear

. set obs 1
Number of observations (_N) was 0, now 1.

. gen long problem = 20120105

. l

     +----------+
     |  problem |
     |----------|
  1. | 20120105 |
     +----------+

. gen year = year(daily(strofreal(problem, "%8.0f"), "YMD"))

. l

     +-----------------+
     |  problem   year |
     |-----------------|
  1. | 20120105   2012 |
     +-----------------+