How can I order the variables of a table of descriptive statistics by their means?

[Andres Santana]

Dear Statalisters,
I have to create many tables of descriptive statistics, and have been asked to order the variables in the tables by their means.
Although it may seem an easy task, I was not able to find a solution.
I will illustrate my problem with auto.dta and just 4 variables:

clear
sysuse auto
su trunk weight length price

Variable | Obs Mean Std. Dev. Min Max
-------------+---------------------------------------------------------
trunk | 74 13.75676 4.277404 5 23
weight | 74 3019.459 777.1936 1760 4840
length | 74 187.9324 22.26634 142 233
price | 74 6165.257 2949.496 3291 15906

I would like to order the variables in, say, ascending order of their means, so as to get something like:

Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------------------
trunk | 74 13.75676 4.277404 5 23
length | 74 187.9324 22.26634 142 233
weight | 74 3019.459 777.1936 1760 4840
price | 74 6165.257 2949.496 3291 15906

I could of course paste the table to Excel and order by "Mean", but I always try to do everything in Stata.
I could of course observe the means obtained and reorder the variables in my command:

su trunk length weight price

But such a solution is quite "homemade", I would very much prefer to be able to automatise it.
Thank you in advanced for your time and your help.
Andrés

[Mike Lacy]

I'm hoping someone else has a more direct solution, but what I thought of was to use -tabstat- to obtain the results and then use -matsort- (from SSC) to sort the results matrix:

tabstat trunk weight length price, save statistics( mean sd min max ) columns(statistics)
// Note that the save option puts the results in the return list,
// but does not obey -columns(statistics)-, hence the transpose
mat Results = r(StatTotal)'
matsort Results 1 up replace

[daniel klein]

I wrote something like this a while ago.

*ssc install vorter
vorter (mean) trunk weight length price , not
summarize `r(varlist)'

yields

. vorter (mean) trunk weight length price , not

. su `r(varlist)'

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
       trunk |        74    13.75676    4.277404          5         23
      length |        74    187.9324    22.26634        142        233
      weight |        74    3019.459    777.1936       1760       4840
      price  |        74    6165.257    2949.496       3291      15906

.
end of do-file

[Andres Santana]

Thank you very much for your two solutions, Mike and Daniel. I would have not arrived at them with my current level of knowledge on Stata (but I keep on learning as much as I can).
Both work very well. For those who are interested in this topic, two clarifications.
First, matsort is not a native command, you can get it in ssc:
ssc install matsort, all
Second, after running the matsort command as Mike explains, to display the matrix you just have to run the following:
matrix list Results
Again, I am very greateful for your help,
Best regards from Madrid,
Andrés

[Bjarte Aagnes]

One could use -svmat- with -frame-, following the tabstat command in #2

sysuse auto , clear

tabstat trunk weight length price, ///
    save statistics( mean sd min max ) ///
    columns(statistics)

matrix stat = r(StatTotal)'  

********************************************************************************

frame create stat
frame change stat

    qui svmat stat , names(col)  // next generate new variable for rownames:

    generate variable = word("`: rowfullnames stat '", _n) if ( _n == _n ) /// 
            , before(`= word("`: colfullnames stat '",  1) ') 

    sort mean
    list

frame change default

********************************************************************************

     +-----------------------------------------------+
     | variable       mean         sd    min     max |
     |-----------------------------------------------|
  1. |    trunk   13.75676   4.277404      5      23 |
  2. |   length   187.9324   22.26634    142     233 |
  3. |   weight   3019.459   777.1935   1760    4840 |
  4. |    price   6165.257   2949.496   3291   15906 |
     +-----------------------------------------------+

[Brandon Olszewski]

daniel klein - WOW! I just found this package you authored and it's so great...especially the <not> option. Much appreciated!