[R] matrix transformation into 3 columns II.

[Roger Bivand]

On Thu, 18 May 2006, Rado Bonk wrote:

> Dear  R-users
> 
> Sorry for posting the previous message too soon before I have finished it.

Hi Rado,

I think reshape() of mprecip as a data frame will do it, or stack(), my 
closest attempts so far were:

reshape(mprecip, direction="long", idvar="date", 
  varying=list(names(mprecip)[2:7]))

which gets "date" right, has the data in the right order, but has a "time" 
variable instead of the station names, or:

stack(mprecip)

which doesn't have the dates, but is otherwise in the expected order. The 
reshape() arguments are a black art ...

Best wishes,

Roger

> 
> I have matrix (mprecip) with headers:
> 
>  > dim(mprecip)
> [1] 6268  170
> 
>  > mprecip
>             date GilzeRijen Eindhoven Volkel ZuidLimburg Arcen Ubachsberg
> 1    01/01/1978         NA       0.0     NA         0.1    NA         NA
> 2    01/02/1978         NA       0.0     NA         0.0    NA         NA
> 3    01/03/1978         NA       1.9     NA         0.7    NA         NA
> 4    01/04/1978         NA       3.5     NA         6.9    NA        6.0
> 5    01/05/1978         NA       1.6     NA         1.8    NA        1.3
> 6    01/06/1978         NA       0.0     NA         0.0    NA         NA
> 7    01/07/1978         NA       0.0     NA         0.0    NA         NA
> 8    01/08/1978         NA       0.0     NA         0.0    NA         NA
> 
> Columns are: DATE and PRECIP values for each station listed in the header.
> 
> I would like to transform the matrix into three columns (database like) 
> to be able to load the data in the database. Here is the output I would 
> like to get get it, number of rows = ncol x nrow. Output should look 
> like this:
> 
> STATION_NAME1 DATE PRECIP
> STATION_NAME1 DATE PRECIP
> STATION_NAME1 DATE PRECIP
> STATION_NAME1 DATE PRECIP
> .
> .
> .
> .
> STATION_NAME2 DATE PRECIP
> STATION_NAME2 DATE PRECIP
> STATION_NAME2 DATE PRECIP
> .
> .
> .
> STATION_NAME170 DATE PRECIP
> STATION_NAME170 DATE PRECIP
> STATION_NAME170 DATE PRECIP
> STATION_NAME170 DATE PRECIP
> STATION_NAME170 DATE PRECIP
> 
> 
> Thanks again for your help.
> 
> 
> Rado
> 
> 

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no