declare data to be panel data

[Rike Lich]

Dear Statalist,
I want this data to be panel data:

* Example generated by -dataex-. To install: ssc install dataex
clear
input int year str397 TITEL long(colors language)
1894 "Erdbeerkuchen"        1 5
1894 "Gedeckte Apfeltorte"  7 7
1896 "Rabarbarkuchen"       3 6
1890 "Blaubertorte"         2 3
1892 "Himbeersorbet"        6 2
1895 "Tiramisu"             5 4
1892 "Avokadocreme"         1 5
1892 "Apfelkuchen"          1 4
1892 "Streuselkuchen"       3 6
1891 "Stachelbeereis"       4 3
1890 "Schokoladenkuchen"    7 3
1880 "Banaenbrot"           7 7
1884 "Pfannkuchen"          3 1
1896 "Ananaskuchen"         8 5
1889 "Eistorte"             2 1
1897 "Baise"                9 5
   . "Linzer Torte"        10 6
   . "Bananensplit"         . 7
end
label values colours colours
label def colours 1 "blau", modify
label def colours 2 "gelb", modify
label def colours 3 "grün", modify
label def colours 4 "himmelblau", modify
label def colours 5 "lila", modify
label def colours 6 "orange", modify
label def colours 7 "rosa", modify
label def colours 8 "rot", modify
label def colours 9 "schwarz", modify
label def colours 10 "weiß", modify
label values language language
label def language 1 "Danish", modify
label def language 2 "English", modify
label def language 3 "French", modify
label def language 4 "German", modify
label def language 5 "Greek", modify
label def language 6 "Italian", modify
label def language 7 "Latin", modify

I already tried:

bysort language colours year : gen n_colours_year = _N
xtset n_colours_year year

and I got the error message:

repeated time values within panel

I would like to have a panel data set about colours and year, that contains the sum of TITELES per language per colour sorted by year.

Thank you very much in advance!!!
Best,
Rike Lich

[Nick Cox]

You asked a similar question on 1 August within https://www.statalist.org/forums/for...moving-average

You can't convince tsset or xtset to define a panel based on identifier and time variable when there are repeated time values.

But the underlying problem is what you need to address: it doesn't obviously make any sense to have a variable containing number of observations as a panel identifier.

A panel structure in Stata is very tightly defined: a numeric identifier, distinct in each panel, and an integer time variable, and at most one observation for each (identifier, time). I can't see that you're even close to that structure.

[Rike Lich]

Dear Nick!

Thank you very much for your remarks!
I solved the problem:

bysort colours year: egen nr_languages = total(languages)
duplicates drop year colours, force
xtset colours year

Best,
Rike Lich