🙋 Equivalent of mat_capp in Mata

[Sergiy Radyakin]

I am looking for the equivalent of mat_capp/mat_rapp commands in Mata.
Specifically, I need to perform the same operations on string matrices which are not possible in the ado-language of Stata.

[Joseph Coveney]

Those two Stata commands match on Stata matrixes' row and column names. Do Mata matrixes have row or column names?

[Niels Henrik Bruun]

You might be interested in the class nhb_mt_labelmatrix which is part of lmatrixtools.(mlib/mata).
The mata/mlib files are the backbone of the packages basetable, matrixtools, and emc.
All of these can be found at the ssc.

The content of lmatrixtools is:

. mata mata describe using lmatrixtools
(library contains 102 members)


      # bytes   type                        name and extent
-------------------------------------------------------------------------------
        1,336   auto classdef scalar        nhb_List()
          248   auto void                    ::append()
          288   auto transmorphic colvector  ::apply()
          212   auto transmorphic colvector  ::content()
          288   auto real matrix             ::find()
          368   auto real colvector          ::frequency()
          244   auto real scalar             ::has_next()
          700   auto transmorphic colvector  ::intersection_unique()
          160   auto real scalar             ::is_empty()
          140   auto real scalar             ::len()
          704   auto transmorphic colvector  ::less_unique()
          212   auto transmorphic scalar     ::next()
          136   auto void                    ::next_init()
          208   auto void                    ::remove()
          192   auto void                    ::reset()
          168   auto string scalar           ::type()
          412   auto transmorphic colvector  ::union_unique()
          212   auto transmorphic colvector  ::unique_values()
          696   auto transmorphic vector    nhb_fp_map()
          432   auto transmorphic vector    nhb_fp_reduce()
        1,256   auto real matrix            nhb_mc_percentiles()
        1,852   auto class scalar           nhb_mc_post_ci_table()
          672   auto real matrix            nhb_mc_predictions()
        1,216   auto real colvector         nhb_mc_smoothed_data()
          416   auto real rowvector         nhb_mc_smoothed_minmax()
          592   auto classdef scalar        nhb_mc_splines()
          216   auto void                    ::add_knots()
          540   auto real matrix             ::marginal_linear()
          156   auto void                    ::new()
          208   auto real vector             ::positive()
        1,200   auto real matrix             ::restricted_cubic()
          484   auto transmorphic matrix    nhb_msa_file_size_kb()
        1,252   auto transmorphic matrix    nhb_msa_oswalk()
          460   auto string scalar          nhb_msa_unab()
        2,288   auto transmorphic matrix    nhb_msa_variable_description()
        1,208   auto classdef scalar        nhb_mt_chi2tabulate()
          544   auto class scalar            ::column_proportions()
          132   auto class scalar            ::counts()
        1,220   auto class scalar            ::counts_with_totals()
          632   auto class scalar            ::expected()
          132   auto class scalar            ::greeks()
        1,048   auto class scalar            ::likelihood_ratio_chisquare_parts()
          380   auto void                    ::new()
        1,040   auto class scalar            ::pearson_chisquare_parts()
          656   auto class scalar            ::proportions()
          544   auto class scalar            ::row_proportions()
        4,964   auto void                    ::set()
          132   auto class scalar            ::tests()
          228   auto void                    ::verbose()
          736   auto transmorphic matrix    nhb_mt_format_real_matrix()
        1,368   auto transmorphic matrix    nhb_mt_justify_string_matrix()
        1,516   auto classdef scalar        nhb_mt_labelmatrix()
          664   auto void                    ::add_sideways()
          664   auto void                    ::append()
          332   auto void                    ::clear()
        1,084   auto transmorphic matrix     ::column_equations()
        1,084   auto transmorphic matrix     ::column_names()
          560   auto string colvector        ::duplicate_strip()
          156   auto real scalar             ::empty()
          596   auto void                    ::from_labelmatrix()
          672   auto void                    ::from_matrix()
          224   auto real matrix             ::hide_small()
          160   auto void                    ::new()
        2,920   auto void                    ::print()
        2,172   auto class scalar            ::regex_select()
        1,084   auto transmorphic matrix     ::row_equations()
        1,084   auto transmorphic matrix     ::row_names()
          784   auto void                    ::to_matrix()
          732   auto class scalar            ::transposed()
          268   auto transmorphic matrix     ::values()
        7,396   auto transmorphic matrix    nhb_mt_mata_string_matrix_styled()
          936   auto transmorphic matrix    nhb_mt_matrix_stripe()
          896   auto transmorphic matrix    nhb_mt_matrix_v_sep()
          736   auto transmorphic matrix    nhb_mt_resize_matrix()
        1,068   auto classdef scalar        nhb_mt_tabulate()
          544   auto class scalar            ::column_proportions()
          132   auto class scalar            ::counts()
        1,220   auto class scalar            ::counts_with_totals()
          632   auto class scalar            ::expected()
        1,048   auto class scalar            ::likelihood_ratio_chisquare_parts()
          220   auto void                    ::new()
        1,040   auto class scalar            ::pearson_chisquare_parts()
          656   auto class scalar            ::proportions()
          544   auto class scalar            ::row_proportions()
        3,112   auto void                    ::set()
        1,888   auto class scalar            ::tests()
          516   auto real colvector          ::validate_varname()
          536   auto string rowvector       nhb_muf_tokensplit()
        1,220   auto transmorphic matrix    nhb_sae_addvars()
        1,636   auto transmorphic matrix    nhb_sae_appendvars()
        1,648   auto real matrix            nhb_sae_collapse()
          964   auto string rowvector       nhb_sae_labelsof()
          472   auto transmorphic matrix    nhb_sae_logstatacode()
          692   auto transmorphic matrix    nhb_sae_markrows()
          244   auto transmorphic matrix    nhb_sae_num_scalar()
        1,612   auto transmorphic matrix    nhb_sae_stored_scalars()
          656   auto transmorphic matrix    nhb_sae_str_mult_matrix()
          884   auto transmorphic matrix    nhb_sae_subselect()
       11,308   auto class scalar           nhb_sae_summary_row()
          848   auto real rowvector         nhb_sae_unique_values()
          932   auto transmorphic matrix    nhb_sae_validate_variable()
          628   auto transmorphic matrix    nhb_sae_variable_data()
-------------------------------------------------------------------------------

If one of the above packages are installed, you can e.g. do:

. sysuse auto
(1978 Automobile Data)

. sumat price mpg if foreign, statistics(mean ci)
-----------------------------------------
                 mean  ci95% lb  ci95% ub
-----------------------------------------
Price         6384.68   5289.07   7480.29
Mileage(mpg)    24.77     22.01     27.54
-----------------------------------------

. matrix foreign = r(sumat)

. sumat price mpg if !foreign, statistics(mean ci)
-----------------------------------------
                 mean  ci95% lb  ci95% ub
-----------------------------------------
Price         6072.42   5230.64   6914.21
Mileage(mpg)    19.83     18.54     21.12
-----------------------------------------

. matrix domestic = r(sumat)

. mata
------------------------------------------------- mata (type end to exit) ---------------------------------------------------------------------
: lmf = nhb_mt_labelmatrix()
: lmf.from_matrix("foreign")
: lmf.print()
-----------------------------------------
                 mean  ci95% lb  ci95% ub
-----------------------------------------
Price         6384.68   5289.07   7480.29
Mileage(mpg)    24.77     22.01     27.54
-----------------------------------------

: lmd = nhb_mt_labelmatrix()
: lmd.from_matrix("domestic")
: lmf.column_equations("Foreign")
: lmd.column_equations("Domestic")
: lmf.add_sideways(lmd)
: lmf.print()
-----------------------------------------------------------------------
              Foreign                      Domestic                    
                 mean  ci95% lb  ci95% ub      mean  ci95% lb  ci95% ub
-----------------------------------------------------------------------
Price         6384.68   5289.07   7480.29   6072.42   5230.64   6914.21
Mileage(mpg)    24.77     22.01     27.54     19.83     18.54     21.12
-----------------------------------------------------------------------

: end
-----------------------------------------------------------------------------------------------------------------------------------------------

I've used -sumat- from package matrixtools to get example matrices.
Note that nhb_mt_labelmatrix also have the method append to put matrices above each other.
Method is not yet public.
Also, I'm quite proud of the method regex_select which can be used to select part of the matrix using regex notation.

Have fun

[Niels Henrik Bruun]

Ah, I did not read your request proporly, sorry.
You want to use string matrices as bases.
So far this is just for matrices of reals with labels, sorry

[Sergiy Radyakin]

Joseph, Neils thank you for responding and sorry about some confusion.

Joseph is right that mata matrices don't have row and column names as Stata matrices do.
I do need to concatenate string matrices because at least partially the content I want to operate on is strings.
My row and column names are just the first row and column of the matrix.
So far I aggregate the data in form of mata string matrices.
The following example illustrates what I am looking for (GOOD)

clear all

mata
    A = "varname", "obs", "state",    "type" \ ///
        "mpg",     "74",  "original", "input" \ ///
        "rep78",   "69",  "modified", "input"

    B = "varname", "obs", "state",    "type" \ ///
        "rep78",   "74",  "original", "input" \ ///
        "dch",     "39",  "original", "calculated"
    A
    B
    BAD=A,B
    "BAD -->"
    BAD
    // Hypothetical command
    // CONCAT_MAT A,B , MATCH("varname") MODE(COLUMNS), EMPTY("~")
    GOOD= "varname", "obs", "state",    "type" , "obs", "state",    "type"\ ///
          "mpg",     "74",  "original", "input", "~",   "~",        "~" \ ///
          "rep78",   "69",  "modified", "input", "74",  "original", "input" \ ///
          "dch",     "~",   "~",        "~",     "39",  "original", "calculated"
    "GOOD -->"
    GOOD
end

So far I can see how this can be done with merging two data files, but I'd like to keep the column titles identical and avoid file operations.

This may be is quite a niche need, so nothing like this may readily exist, but just wanted to check if there is a ready-made solution before writing my own.

Thank you again, Sergiy

[Sergiy Radyakin]

The following post described a related task:
https://theesspreckelsen.wordpress.com/2014/07/

I have by now solved my problem with a custom code, so this thread can be closed.