Using for each for observations in stata

[George Paily]

For each observations, variable x takes the value 1 in year 2000, how to make the value of x =1 in subsequent years ?

Any help is greatly appreciated

[Nick Cox]

You're not telling us much about your data. I am going to guess at some kind of panel or longitudinal layout with an identifier as well as a year variable.

gen wanted = x == 1 & year == 2000 
bysort id (year) : replace x = 1 if x[_n-1] == 1

Or

bysort id (year) : gen wanted = sum(x == 1 & year == 2000)

[Rich Goldstein]

without a -dataex- example (see the FAQ), it is not possible to give exact code but my best guess is that, after sorting, you just want something like:

replace x=1 if year>2000

but note that this makes a number of assumptions that may not be true

note that it is very unlikely that loop is needed here

[George Paily]

Thanks Nick and Rich> Mine is an unbalanced panel data with firm_id and year.

r new_product_dummy

. rename new_product_dummy x

. dataex Factory_Code year x

----------------------- copy starting from the next line -----------------------

* Example generated by -dataex-. For more info, type help dataex
clear
input str8 Factory_Code float(year x)
"0000003F" 2000 0
"0000003F" 2002 1
"0000006F" 2010 0
"0000009F" 2001 0
"0000028F" 2011 0
"0000033F" 2005 1
"0000033F" 1999 0
"0000033F" 2001 0
"0000033F" 2000 0
"0000033F" 2002 0
"0000033F" 2010 1
"0000103F" 2004 1
"0000103F" 2011 1
"0000103F" 2003 1
"0000103F" 2000 0
"0000106F" 2015 0
"0000106F" 2014 1
"0000108B" 2006 0
"0000108B" 2015 0
"0000108B" 2013 1
"0000110F" 2013 0
"0000129F" 2000 0
"0000129F" 2005 1
"0000133F" 2003 0
"0000133F" 2015 1
"0000133F" 2009 1
"0000133F" 2010 0
"0000206F" 2009 1
"0000206F" 2011 1
"0000206F" 2010 1
"0000206F" 2015 0
"0000206F" 2016 1
"0000206F" 2013 1
"0000206F" 2008 1
"0000206F" 2014 0
"0000208B" 2009 1
"0000208B" 2006 0
"0000208B" 2012 0
"0000208B" 2011 1
"0000208B" 2010 0
"0000208B" 2007 0
"0000208B" 2008 0
"0000208B" 2013 0
"0000209F" 2005 0
"0000209F" 2013 1
"0000209F" 2003 0
"0000219F" 2004 0
"0000220F" 2014 0
"0000220F" 2015 0
"0000233F" 2005 0
"0000233F" 2009 1
"0000233F" 2011 1
"0000306F" 2008 0
"0000306F" 2007 0
"0000306F" 2014 1
"0000308B" 2011 1
"0000308B" 2007 0
"0000308B" 2012 0
"0000319F" 2005 1
"0000319F" 2004 0
"0000333F" 2006 0
"0000333F" 2011 1
"0000333F" 2013 1
"0000406F" 2008 1
"0000406F" 2007 0
"0000408B" 2006 0
"0000408B" 2010 1
"0000408B" 2016 1
"0000419F" 2004 0
"0000419F" 2005 1
"0000420F" 2011 0
"0000424F" 2003 1
"0000424F" 1999 0
"0000424F" 2007 0
"0000424F" 2001 1
"0000429F" 2001 0
"0000429F" 2013 0
"0000429F" 2011 1
"0000503F" 2004 0
"0000503F" 2000 0
"0000503F" 2015 1
"0000519F" 2004 0
"0000519F" 2013 1
"0000520F" 2011 0
"0000523F" 2015 0
"0000527F" 2006 0
"0000527F" 2009 0
"0000527F" 2011 1
"0000527F" 2003 0
"0000533F" 2008 1
"0000533F" 2001 0
"0000533F" 2000 0
"0000603F" 2006 0
"0000609F" 2007 1
"0000609F" 2004 0
"0000610F" 2011 0
"0000619F" 2003 1
"0000619F" 2002 0
"0000619F" 2004 1
"0000620F" 2015 0
end

[Nick Cox]

Some variation on #2 remains my guess. Whether it's firm identifier or factory code you care about only you can say.