[Jean-Michel Galarneau]

graph combine to work with meta forest without having to move everything to the user-written metan suite.

[Todd Jones]

#344 yes, good point.

[FernandoRios]

Something I keep finding as a problem with other colleagues.
Would it be possible to allow Stata to "read" ster files automatically?
I know one could just "est use" it. But interestingly enough, many people do not know how to.
If one could double click it so it runs a simple "ereturn display" would be very useful, and perhaps not hard to implement

[Anna Stubbendorff]

It would be great to have the possibility to show row percentages in dtable instead of column percentages.

[Luis Pecht]

on #349, Andrew Musau teached how to show row percentages,

https://www.statalist.org/forums/for...-to-each-other

sysuse auto, clear
collect clear
table (foreign), nformat(%3.2f ) stat(percent) stat( mean weight) nototals name(c1)
table (rep78), nformat(%3.2f ) stat(percent) stat( mean weight) nototals name(c2)
collect combine all= c1 c2
collect levelsof across
collect layout (`s(levels)') (var#result)

. collect layout (`s(levels)') (var#result)
(dimension _hide not found)

Collection: all
      Rows: _hide foreign rep78
   Columns: var#result
   Table 1: 9 x 2

---------------------------------------------
                   |  Weight (lbs.)   Percent
                   |           Mean          
-------------------+-------------------------
Car origin         |                        
  Domestic         |        3317.12     70.27
  Foreign          |        2315.91     29.73
Repair record 1978 |                        
  1                |        3100.00      2.90
  2                |        3353.75     11.59
  3                |        3299.00     43.48
  4                |        2870.00     26.09
  5                |        2322.73     15.94
---------------------------------------------

plus Bjarte Aagnes, alternative:

   
collect : table (foreign), nformat(%3.2f ) stat(percent) stat( mean weight) nototals
collect : table (rep78), nformat(%3.2f ) stat(percent) stat( mean weight) nototals append
collect layout (foreign rep78)(result)

      Rows: foreign rep78
   Columns: result
   Table 1: 9 x 2

---------------------------------------
                   |  Percent      Mean
-------------------+-------------------
Car origin         |                  
  Domestic         |    70.27   3317.12
  Foreign          |    29.73   2315.91
Repair record 1978 |                  
  1                |     2.90   3100.00
  2                |    11.59   3353.75
  3                |    43.48   3299.00
  4                |    26.09   2870.00
  5                |    15.94   2322.73
---------------------------------------

[Tobias Bergsmann]

'Highlight Occurrences' or 'Word Highlighting' would be nice. Like in VS Code, which has it per default: when i select a word (e.g. a macro name or whatever) all occurences of the same word are also highlighted in the same code file in the editor, which is incredible usefull while editing code.
Example:

[Clyde Schechter]

Re #350: It is true that this can be done in the ways shown. But the -collect- command is very complicated and learning to use it effectively requires a great deal of investment of time and effort. -dtable- is a convenience command that is intended to produce a commonly needed table with a single command. It would be even more convenient if it had an option to show row percentages.

[Stephen Jenkins]

-regress- does not currently store the sum of the weights in e() when weights are used but the information is clearly available (it is echoed to output as the command runs). Please can this sum be added to e()! For further duscussion, please see https://www.statalist.org/forums/for...-after-regress

[Luis Pecht]

I wish table command had the number of unique elements statistics, same way gcollapse (nunique) does.

[Elizabeth Anderson]

I would like to see FIML estimation in the gsem function.

[Wah Myint]

Please excuse me if someone already mentioned it. I wish Stata 19 has (a) SEM for dichotomous or categorical variable ---somethink like it can easily produce in Mplus, (b) for social network analysis, and (3) data import to/export from .dat.

[Henry Strawforrd]

I will wish every year the same: https://pola.rs/

[Daniel Schaefer]

Wah Myint Regarding (a), I believe GSEM already supports these variables types. Regarding (c), the .dat file extension is just a generic "data" file extension. .dat files are not in a specific file format. The format of a .dat file depends entirely on the software that defines the .dat file. There is not a generic way to import/export these files.

[Erik Ruzek]

#356 Wah Myint - Stata has the gsem command for structural equation models with binary, ordinal, nominal, and count outcomes. gsem can also do multilevel SEM as well as latent profile and latent class analyses. It's not as fast as Mplus, and it cannot do everything Mplus can do, but it is a good enough solution for many problems.

Edit: Sorry, i see that Daniel addressed this. I added a little more context.

[Niels Henrik Bruun]

When using user-written commands, ancillary files are saved in the current folder.
Later, you might not be able to find these files. It would be better if the ancillary were saved in the personal folder.
This way You get datasets into Stata simply by -use, ancillary_dataset.dta, clear-.