[R] how to merge panel data stored by variable?

Viktor Nagy viktor.nagy at gmail.com
Thu Dec 11 01:40:17 CET 2008


wow, this package is amazing! Thank you!
V

2008/12/10 hadley wickham <h.wickham at gmail.com>:
> On Wed, Dec 10, 2008 at 1:54 PM, Viktor Nagy <viktor.nagy at gmail.com> wrote:
>> 2008/12/10 Stefan Grosse <singularitaet at gmx.net>
>>>
>>> > I have two datasets stored in tab-separated format in the following way
>>> > file1:
>>> > country           year1        year2
>>> > Germany        var1          var1
>>> > Hungary         var1          var1
>>> >
>>> > file2:
>>> > country           year1        year2
>>> > Germany        var2         var2
>>> > Hungary         var2         var2
>>> >
>>> > I can easily read in these files, but how can I merge them as a panel
>>> > dataset?
>>> >
>>> >
>>>
>>> ?cbind
>>> ?rbind
>>>
>>
>> Thanks for your answer Stefan, but I think cbind/rbind does not solve
>> my problem.
>>
>> After reading ?cbind I realised that first I should convert all the
>> dataset to panel separately, and then I will be able to cbind/rbind
>> it.
>>
>> So the question is how to read in the following data from a file as a
>> panel dataset
>>
>> country           year1        year2
>> Germany        var1          var1
>> Hungary         var1          var1
>>
>> that is to have:
>> country        year      var1
>> Germany      year1    value1
>> Germany      year2    value1
>> Hungary       year1    value1
>> Hungary       year2    value1
>
> Have a look at melt in the reshape package, http://had.co.nz/reshape.
>
> melt(mydf, id = "country")
>
> should do what you want.
>
> Hadley
>
> --
> http://had.co.nz/
>



-- 
Viktor Nagy - http://viktornagy.com
PhD student
Toulouse School of Economics



More information about the R-help mailing list