dropping observations if variable is zero between time period

[Thomas Kiekebosch]

Hello,

Could someone please help me dropping some observations? I'm investigating the influence of patenting on the return on investment of firms. Now I want to drop the companies that never had a patent between variable fyear (fiscal year) 1997-2015 (see the example company in the screenshot). But I want to keep them if they had a patent in this time period. So, also if they only have 1 patent in that time period, I want to keep them in my dataset. I'm struggling to find the right code so could someone please help me?

Thanks!

[Thomas Kiekebosch]

So in this case companies like RELM WIRELESS CORP i want to drop

[William Lisowski]

Something like the following might do what you want.

by gvkey (fyear), sort: egen tokeep = max(cond(inrange(fyear,1997,2015),pat_year,0)
drop if tokeep==0

To understand the code,

help cond()
help egen
help by

[Thomas Kiekebosch]

Thank you for responding. Now I'm getting a error called 'parentheses unbalanced' r132. I tried a few thing but couldn't make it work. Sorry I'm really a nood in stata

[William Lisowski]

Here is how you diagnose the source of unbalanced parentheses —the fact that there are three left parentheses but just two right parentheses.

max(cond(inrange(fyear,1997,2015),pat_year,0))
                |_______________|

Does help inrange() tell us this is the correct syntax for the inrange() function?
Yes.

max(cond(inrange(fyear,1997,2015),pat_year,0)
                |_______________|
        |___________________________________|

Does help cond() tell us this is the correct syntax for the cond() function?
Yes.

max(cond(inrange(fyear,1997,2015),pat_year,0)
                |_______________|
        |___________________________________|
   |__________________________________________ no matching )

So the max() function is missing its right parenthesis.

by gvkey (fyear), sort: egen tokeep = max(cond(inrange(fyear,1997,2015),pat_year,0))
drop if tokeep==0

[Thomas Kiekebosch]

Thank you for your clear response! I understand it now. The code worked!