[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: generating new observations
That's one interpretation of what is wanted here.
Another is that this question is asking for ad hoc surgery to fix
a specific problem, the fact that for trade purposes
Belgium and Luxembourg are a unit.
We also have, on this evidence, a consistency problem,
as "Luxemburg" and "Luxembourg" are two and the same
thing. (We professional geographers are well up on such
I'd do ad hoc surgery this way:
gen byte tojoin = inlist(importer, "Belgium", "Luxembourg", "Luxemburg")
After this, these countries code as 1 and everything else as 0.
If you also have "Belgie", "Belgique" or any other variations, add
them to the parenthesised list.
Now we want to add up commensurables:
bysort tojoin commoditycode : replace tradevalue = sum(tradevalue) if tojoin
and possibly to clean up:
by tojoin commoditycode : keep if _n == _N & tojoin
But -save- the dataset first, as this may not be what you want.
Also, isn't there a parallel issue with exporters?
> At 12:45 PM 3/29/2006, Kremena wrote:
> >I have a data set with the following variables:
> >importer exporter commoditycode tradevalue
> >Belgium Argentina 1 123
> >Belgium Argentina 2 456
> >Luxemburg Argentina 1 789
> >Luxembourg Argentina 3 147
> >I want to generate new observations such that importer is
> >and tradevalue is the combined trade value for the two
> countries for the
> >respective commodity.
> >Can someone tell me how I can do that?
> You can do...
> collapse (sum) tradevalue , by(commoditycode)
> This will sum to the level of commoditycode, regardless of importer
> or exporter (and it won't create a combined importer identifier). If
> you want specific groups of importers or importer/exporter
> combinations, you will need to do that in advance, creating another
> variable, and then put that variable into the by() option.
* For searches and help try: