summing observations by id

[Stef Luck]

Hi all,

I am sorry if I this may be a trivial question and if it has already been discussed. I searched through the forum but didn't find a solution.

I have bilateral data on migration and I would like to create a new variable in which the observation of Belgium and Luxembourg are summed, obtaining a "Belgium-Luxembourg" i_country (keeping all other countries and variables.

The sum (of Total, Male and Female) should be by year, destination country (j_*).

Could please someone give me some advice??

Hoping that I have made this clear, thank you in advance!

i_country	i_code	i_iso2	i_iso3	j_country	j_code	j_iso2	j_iso3	year	Total	Male	Female
Luxembourg	442	LU	LUX	United Arab Emirates	784	AE	ARE	2017
Belgium	56	BE	BEL	United Arab Emirates	784	AE	ARE	2017
Belgium	56	BE	BEL	United Kingdom	826	GB	GBR	2017	36822	16373	20449
Luxembourg	442	LU	LUX	United Kingdom	826	GB	GBR	2017	1714	793	921
Belgium	56	BE	BEL	United Republic of Tanzania	834	TZ	TZA	2017
Luxembourg	442	LU	LUX	United Republic of Tanzania	834	TZ	TZA	2017
Belgium	56	BE	BEL	United States Virgin Islands	850	VI	VIR	2017
Luxembourg	442	LU	LUX	United States Virgin Islands	850	VI	VIR	2017
Belgium	56	BE	BEL	United States of America	840	US	USA	2017	36310	17775	18535
Luxembourg	442	LU	LUX	United States of America	840	US	USA	2017	2626	1290	1336
Belgium	56	BE	BEL	Uruguay	858	UY	URY	2017
Luxembourg	442	LU	LUX	Uruguay	858	UY	URY	2017

[Carlo Lazzaro]

Stef:
minor surgery here would imply to export in another .dta the observations concerning Belgium and Luxemburg, -collapse- appropriately and then -append- back to the original file.

[Marcos Almeida]

Carlo gave excellent advice. Please read the FAQ. There you'll find how to share data\command\output. This is the best approach to entail the precise code you're looking for.

[Stef Luck]

Dear Carlo and Marcos,
thank you for your reply.
I think I would go with Carlo's suggestion.

Thank you again