labels on mata matrices

[Niels Henrik Bruun]

Hi
One of the things I like about the -matrix- in Stata is the possiblity to add row and column labels to matrices there.
Is there something similar for Mata matrices?
I've been trying to find something in the documentation without luck, so the answer could be that it is not possible.
But in that case what do others do if they would like to use this facility?

Looking forward to hear from you

[Jeff Pitblado (StataCorp)]

There is no built-in facility for row and column labels for Mata matrices.

However, you can build separate string matrices that correspond to the
row and column labels (also known as matrix stripes) for any given
matrix. I image one trouble you have is trying to display your Mata
matrix using the labels you have built.

There is an not-documented Mata function _matrix_list() that was
developed to do this for optimize() and moptimize(). The
source for _matrix_list() is in your adopath.

viewsource st_matrix_list.mata

In fact, this is the code produces the output for the
matrix list command.

The basic syntax for _matrix_list() is

_matrix_list(matrix, rstripe, cstripe)

matrix is a numeric Mata matrix.
rstripe is a string Mata matrix representing the row labels.
cstripe is a string Mata matrix representing the column labels.

The rstripe and cstripe matrices both have 2 columns.
Column 1 is the equation part, and Column 2 is the coefficient part.
The rows of rstripe correspond with the rows of matrix.
The rows of cstripe correspond with the columns of matrix.

Here is a quick example to see that _matrix_list() will produce the
same output as matrix list.

sysuse auto
gsem mpg <- turn i.rep
matrix list e(b)

mata:
b = st_matrix("e(b)")
rlabels = st_matrixrowstripe("e(b)")
clabels = st_matrixcolstripe("e(b)")
rlabels
clabels
_matrix_list(b,rlabels,clabels)
end

Here is a log.

. sysuse auto
(1978 Automobile Data)

. gsem mpg <- turn i.rep

Iteration 0:   log likelihood = -189.61577  
Iteration 1:   log likelihood = -189.61577  

Generalized structural equation model             Number of obs   =         69
Log likelihood = -189.61577

------------------------------------------------------------------------------
             |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
mpg <-       |
        turn |  -.9131143   .1214961    -7.52   0.000    -1.151242   -.6749863
             |
       rep78 |
          2  |   .2936464   3.000444     0.10   0.922    -5.587115    6.174408
          3  |  -1.505792   2.758856    -0.55   0.585    -6.913051    3.901466
          4  |  -1.616119   2.832069    -0.57   0.568    -7.166871    3.934633
          5  |   1.466023   2.976163     0.49   0.622    -4.367149    7.299196
             |
       _cons |   58.43769   5.652368    10.34   0.000     47.35925    69.51612
-------------+----------------------------------------------------------------
   var(e.mpg)|   14.27104   2.429663                      10.22204    19.92387
------------------------------------------------------------------------------

. matrix list e(b)

e(b)[1,8]
            mpg:         mpg:         mpg:         mpg:         mpg:
                          1b.           2.           3.           4.
           turn        rep78        rep78        rep78        rep78
y1   -.91311427            0     .2936464   -1.5057924    -1.616119

            mpg:         mpg:  var(e.mpg):
              5.                          
          rep78        _cons        _cons
y1    1.4660234    58.437685    14.271038

. 
. mata:
------------------------------------------------- mata (type end to exit) -----
: b = st_matrix("e(b)")

: rlabels = st_matrixrowstripe("e(b)")

: clabels = st_matrixcolstripe("e(b)")

: rlabels
        1    2
    +-----------+
  1 |       y1  |
    +-----------+

: clabels
                1            2
    +---------------------------+
  1 |         mpg         turn  |
  2 |         mpg     1b.rep78  |
  3 |         mpg      2.rep78  |
  4 |         mpg      3.rep78  |
  5 |         mpg      4.rep78  |
  6 |         mpg      5.rep78  |
  7 |         mpg        _cons  |
  8 |  var(e.mpg)        _cons  |
    +---------------------------+

: _matrix_list(b,rlabels,clabels)
            mpg         mpg         mpg         mpg         mpg         mpg
          turn    1b.rep78     2.rep78     3.rep78     4.rep78     5.rep78
y1  -.91311427           0    .2936464  -1.5057924   -1.616119   1.4660234

            mpg  var(e.mpg)
         _cons       _cons
y1   58.437685   14.271038

: end

[Niels Henrik Bruun]

Hi Steve
Thank you very much.
I'll look into it