Export file if conditions are observed

[Nicolas Charette]

Hello,

I am using Stata/BE 17 for Mac and I want to export an Excel file depending on a merge result.

The code below works when merge==1 is observed, but when it is not, of course all observations are deleted and no n=_n is generated, and then command keep if n==1 doesn't work (n ambiguous abbreviation).

keep if _merge==1
bysort id: gen n=_n
keep if n==1
keep id lname fname
if c(N) > 0 {
export excel $orig/questions.xlsx, replace firstrow(var)
}

Could someone help me, please? I want the code to work even if merge==1 is not observed.

Thank you,

Nicolas

[Clyde Schechter]

Well, what do you want the code to actually do if there are no observations with _merge == 1?

[Hemanshu Kumar]

In addition to answering the question in #2, please also provide an example of your data after merging (using dataex), when _merge == 1 is not observed.

[Nicolas Charette]

Clyde Schechter I use this part of the code between a "preserve" and "restore" segment in a much larger do-file. Additional data will come in regularly later in the process, and I use this part of the code to flag certain problems in the data gathering process when there are observations for _merge==1. If there are no observations with _merge==1, it means there are no problems and I just want the code to run smoothly and not stop.

[Nicolas Charette]

Hemanshu Kumar

Here is a shorter version of the data. I have erased lname and fname for anonymity.

* Example generated by -dataex-. For more info, type help dataex
clear
input long id int annee byte(note _merge)
8198541 2011 20 3
8198541 2016 48 3
8198541 2011 33 3
8198541 2019 56 3
8198541 2019 79 3
8198541 2011 20 3
8198541 2016 23 3
8198541 2012 40 3
8198541 2015 83 3
8206906 2017 68 3
8206906 2012 78 3
8206906 2016 59 3
8206906 2016 56 3
8206906 2015 50 3
8206906 2017 66 3
8206906 2015 78 3
8206906 2014 52 3
8206906 2011 52 3
8206906 2016 55 3
8206906 2016 80 3
8206906 2016 67 3
8206906 2015 80 3
8206906 2012 77 3
8206906 2012 64 3
8206906 2009 84 3
8206906 2009 77 3
8206906 2013 73 3
8206906 2017 67 3
8206906 2009 73 3
8206906 2017 72 3
8206906 2009 80 3
8206906 2013 50 3
8206906 2014 82 3
8206906 2015 79 3
8206906 2011 85 3
8206906 2012 51 3
8206906 2016 53 3
8206906 2016 72 3
8206906 2012 60 3
8206906 2011 62 3
8206906 2016 54 3
8206906 2009 61 3
8206906 2017 60 3
8206906 2015 76 3
8206906 2016 71 3
8206906 2013 60 3
8206906 2009 55 3
8206906 2011 65 3
end
label values _merge _merge
label def _merge 3 "Matched (3)", modify

[Nicolas Charette]

With that added info from dataex and my comment (#4), what I wrote in my initial post would look more like:

preserve
keep if _merge==1
bysort id: gen n=_n
keep if n==1
keep id  
if c(N) > 0 {
       export excel $orig/questions.xlsx, replace firstrow(var)
}
restore

My intuition is that I would have to code something that starts with "if merge==1 is observed" kind of command...

[Hemanshu Kumar]

This should do it:

preserve
    count if _merge == 1
    if `r(N)' {
        keep if _merge==1
        bysort id: gen n=_n
        keep if n==1
        keep id  
        if c(N) > 0 {
               export excel $orig/questions.xlsx, replace firstrow(var)
        }
    }
restore

[Nicolas Charette]

Thank you ! It worked fine.

N.