Removing first numbers on a numeric variable

[Rita Neves]

Hi!

I have been looking around for a simple way to remove part of my numeric variable, but didn't find an answer. Could anyone help, please?
Current variable (yearmonth -numeric) --> 200101 (all starting with 200)
What I want to keep ---> 101
So I want to remove the first bit, 200, but maintain it as a number and ideally the name as well.
Thank!!

[Rich Goldstein]

if all values are similar to the one you show you can use the following:

gen newvar = real(substr(string(200101),4,3)

replace "newvar" with your choice

replace "200101" with the variable name in your data set

[daniel klein]

generate newvar = yearmonth-200000

Best
Daniel

[Rita Neves]

Of course, it makes total sense what you mention, Daniel!
I will also try that one, Rich
Thanks for helping me out!
Accidentally I got to something like:
replace yearmonth = mod( yearmonth, 1000)
It does seem to work!

[Nick Cox]

Although people answered your question as you asked, if your variable contains values like 200101, 200102, 200112, 200201, 200202, ... then you have solved one problem only to create another as values like 101, 102, ..., 112, 201, 202,.... are just as awkward. This problem is discussed in a Tip to appear in Stata Journal 18(3). Meanwhile the best way to get a monthly date out of your date is something like this

 clear 
 input yearmonth 
 200101 
 200102
 200112 
 200201
 200202
 end 
 
 gen mdate = ym(floor(yearmonth/100), mod(yearmonth, 100)) 
 
 format mdate %tm 
 
 list
 
      +--------------------+
     | yearmo~h     mdate |
     |--------------------|
  1. |   200101    2001m1 |
  2. |   200102    2001m2 |
  3. |   200112   2001m12 |
  4. |   200201    2002m1 |
  5. |   200202    2002m2 |
     +--------------------+

This also works with numdate (SSC): (see e.g. https://www.statalist.org/forums/for...date-variables)

numdate monthly mdate2 = yearmonth, pattern(YM)