Re: st: How to create a blank n x n matrix; how to refer to variables without using ID; how to extract colnames and rownames from the dataset.

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Wed, 9/3/11, Chen, Xianwen wrote:
> The first question is a fast way to create a n x n matrix.
> For example, to create a 8 x 8 matrix for later data
> manipulation.

matrix foo = J(8,8,.)

This creates a 8x8 matrix containing all missing values (.).

> The second question is regarding variables. I want to
> search through the variables by each entry of the data. I
> can specify the column IDs but I want a more automated
> solution.
> 
> Is there a way of working with it, say somefunction{1,}
> will actually refer to the first column of the dataset?
 
In Stata you almost always work with variable names and not
column numbers. It really really really pays to start learning
that if you want to be an effective user of Stata.

Having said that you can do it:

ds
local varl "`r(varlist)'"

first variable (column) is called `: word 1 of `varl''
second variable is called `: word 2 of `varl'' 
etc.
 
> The third question is to extract colnames and rownames from
> the datasets. Is there a function I can use?

column names are variable names in Stata. You can get that
using -ds-

rownames are just observation numbers, you can get those using
_n

It looks to me that you are comming from another package that
works primary with matrices. Stata does not function that way,
in the sense that it primarily works on a dataset with 
variables and observations instead of rows and columns. All 
functions are geared towards that structure. You can try to 
translate that to matrix manipulations, but then you will be 
making your live unnecesarily hard (and your computations 
unnecesarily slow) because Stata was just not designed for 
that purpose. If you do not want to change your habbits then 
your best bet would be to use Mata instead of Stata, because 
Mata is designed for dealing with matrices instead of dealing 
with data. 

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/