labelling rows and columns of summary statistics with esttab

[Sebastien Willis]

I am producing tables of summary statistics with -esttab-, from -ssc-, in Stata 14. Each statistic is in its own row, like in the following:

sysuse auto, clear
estimates clear
eststo: estpost su mpg
esttab, cells(mean(fmt(2)) sd(fmt(2)) count(label(N) fmt(0)))

The output table looks like this:

-------------------------
                      (1)
                        
                mean/sd/N
-------------------------
mpg                 21.30
                     5.79
                       74
-------------------------
N                      74
-------------------------

Question 1: How do I swap the variable name, mpg, and the names of the statistics, so that furthermore each statistic name is in its own row?
Question 2: How do I make the column number appear below all other model or group titles or column labels? I found this analogous question on statalist, but seemingly no answer.

My ideal table would look something like this:

-------------------------
                        
                     mpg
                      (1)
-------------------------
mean           21.30
SD                 5.79
N                       74
-------------------------
N                      74
-------------------------

Thanks for all your help.

[Simon Walker]

Did you find a solution?

[Frode Andre]

Is this what you are looking for?

esttab, cells(mean(fmt(2)) sd(fmt(2)) count(label(N) fmt(0))) label replace

[Frode Andre]

I am sorry, the code in #3 is not what you were looking for. I posted prematurely (and it takes forever to wait for editing the post).

[Andrew Musau]

This thread was started nearly five years ago, but the most recent response was just a few days ago. Here, you can define a new matrix for the results to achieve the intended format in #1. Use of -collab()- and -mlab()- can fix the numbering issue. -noobs- eliminates the duplication of observation counts.

sysuse auto, clear
estimates clear
eststo m1: estpost su mpg
mat wanted= e(mean)\ e(sd)\ e(count)
mat rownames wanted = mean sd N
estadd mat wanted= wanted': m1
esttab, cells(wanted) mlab(mpg) collab((1)) nonumb noobs

Res.:

. esttab, cells(wanted) mlab(mpg) collab((1)) nonumb noobs

-------------------------
                      mpg
                      (1)
-------------------------
mean              21.2973
sd               5.785503
N                      74
-------------------------

[Sebastien Willis]

Thanks for the responses, I don't even remember what I ended up doing at the time, but Andrew's solution certainly delivers what I was asking for.