Identifying unmerged observations by group

[Scott Rick]

Hi. In the data below, I have merged two datasets by a common code variable to have the variables Product, year, and merge. I want to identify if there are Product names and years that have no match. Basically a flag for a Product-year group within which there is no _merge==3 value at all. For example, in the data below, the last five observations of ACIPHEX in 2008 should be flagged. Any help will be much appreciated.

input str35 Product float year1 byte _merge
"ACANYA"   2009 1
"ACANYA"   2009 3
"ACANYA"   2010 1
"ACANYA"   2010 3
"ACANYA"   2011 1
"ACANYA"   2011 3
"ACANYA"   2012 1
"ACANYA"   2012 3
"ACCOLATE" 2007 1
"ACCOLATE" 2007 1
"ACCOLATE" 2007 1
"ACCOLATE" 2007 1
"ACCOLATE" 2007 1
"ACCOLATE" 2007 1
"ACCOLATE" 2007 1
"ACCOLATE" 2007 1
"ACCOLATE" 2007 1
"ACCOLATE" 2007 3
"ACIPHEX"  2007 1
"ACIPHEX"  2007 1
"ACIPHEX"  2007 1
"ACIPHEX"  2007 3
"ACIPHEX"  2007 3
"ACIPHEX"  2007 3
"ACIPHEX"  2008 1
"ACIPHEX"  2008 1
"ACIPHEX"  2008 1
"ACIPHEX"  2008 1
"ACIPHEX"  2008 1

[William Lisowski]

Perhaps this will start you in a useful direction.

egen wanted = min(_merge!=3), by(Product year1)

[Scott Rick]

Thanks, William. That worked perfectly as always