grouping observations with specific characteristics

Hi all,
I have the following table in excel
time country Location price2
1/09 Austria l1 12
2/09 Austria l1 4
3/09 Austria l1 2
1/09 Austria l2 1
2/09 Austria l2 23
3/09 Austria l2 14
1/09 Germany l1 100
2/09 Germany l1 110
3/09 Germany l1 120
1/09 Germany l2 100
2/09 Germany l2 110
3/09 Germany l2 120
I want to obtain the following structure
1/09 Austria l1 12
1/09 Germany l1 100
2/09 Austria l1 4
2/09 Germany l1 110
3/09 Austria l1 2
3/09 Germany l1 120
1/09 Austria l2 1
1/09 Germany l2 100
2/09 Austria l2 23
2/09 Germany l2 110
and so forth...
the first 3 varibles are string varibles.
Is there any "qick code" for this as i have a large matrix?
Thanks in advance,

1 comentario

Oleg Komarov
Oleg Komarov el 21 de Mayo de 2012
This post has been duplicated on: http://www.mathworks.com/matlabcentral/answers/38955-conveting-a-string-of-variables-to-numbers

Iniciar sesión para comentar.

 Respuesta aceptada

Geoff
Geoff el 21 de Mayo de 2012
You mean you want to sort by Location, then Time, then Country, then Price?
Before you go MatLabbing this... Would you consider that you can do this very easily in Excel?
Now, if you want to do it in MatLab, read all your data into an N-by-4 cell array (cell matrix, I guess), then call sortrows with your column order:
sortedData = sortrows( origData, [3, 1, 2, 4] );

4 comentarios

Geoff
Geoff el 21 de Mayo de 2012
Oh yes, and if your dates are not immediately sortable as a string, you will need to either modify them to be sortable, or convert them to a value that is.
ektor
ektor el 21 de Mayo de 2012
Hi Geoff,
I want to gather together first all l1s that correspond to 1/09
then, i want to gather all l1s that correspond to 2/09
then, i want to gather all l1s that correspond to 3/09
then, i want to gather all l2s that correspond to `1/09
then, i want to gather all l2s that correspond to `2/09
then, i want to gather all l2s that correspond to `3/09
and so forth.....
Thanks
Geoff
Geoff el 21 de Mayo de 2012
Like I said. Sort by location and date (others are optional). If you can have I10 or more, or dates 10/09 or more, then you need to convert those... You could convert the 'I' column with something like this - might need syntax tweaks:
data{:,3} = cellfun( @(x) str2double(x(2:end)), {data{:,3}});
For the dates, you could pad a leading zero onto anything with only 4 characters.
ektor
ektor el 22 de Mayo de 2012
thank you Geoff!

Iniciar sesión para comentar.

Más respuestas (0)

Categorías

Preguntada:

el 21 de Mayo de 2012

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by