[Ryckie Wade]

Support for pooling proportions in the meta suite (e.g. incorporation of Marc Arbyn's metaprop), please

[Clyde Schechter]

It would be convenient to have a new macro list function that introduces commas between tokens in a macro. It would work much like the -separate()- option in -levelsof-, but would be applicable to any macro. While comma-separated lists are not all that common in Stata, it is sometimes necessary to take, say, a list of variables that one wants to iterate over, or feed as a varlist to some command, and also pass it as a list of arguments to a function (e.g. -missing()-). The insertion of commas into such lists can be ticklish at times if there are tokens that contain embedded blanks, or extra leading or trailing blanks. It would be great not to have to deal with that kind of low-level detail in code.

[Leonardo Guizzetti]

Support for pooling proportions in the meta suite (e.g. incorporation of Marc Arbyn's metaprop), please

With a little work to calculate your own effect size and standard error (e.g., the logit of the proportion and the standard error), you can feed this into the meta suite. The same applies if you had a single group mean, or whatever else.

[William Lisowski]

One last off-topic post to close the discussion regarding community contributed packages hosted at UCLA, begun above at #289.

Thanks to the help of our friends at StataCorp, the packages now hosted at UCLA in https://stats.oarc.ucla.edu/stat/stata/ado/analysis are included in the nightly search indexing. Packages can be installed in the usual fashion of clicking on the package name in the search results and then clicking on "click here to install". Anyone experiencing the results like the following from ado update

   [20] fstar at http://www.ats.ucla.edu/stat/stata/ado/analysis:
      server not responding or package is no longer available

can resolve the issue by using ado uninstall to remove the copy installed from the defunct URL and then reinstall the package as described above or using

. net install fstar, from(https://stats.oarc.ucla.edu/stat/stata/ado/analysis)
checking fstar consistency and verifying not already installed...
installing into /Users/lisowskiw/Library/Application Support/Stata/ado/plus/...
installation complete.

.

[John Mullahy]

Re: Clyde Schechter in #302: I'm not 100% sure I'm understanding but does this accomplish what you want:

loc t="This is a long string"
mata st_local("sc",invtokens(tokens("`t'"),","))
di "`sc'"

[William Lisowski]

Regarding inserting commas between tokens in a macro value:

Clyde in #302 specifically mentioned tokens containing embedded blanks, such as the following example, which breaks the sample code in #305, although that may be just a matter of getting compound quotation marks set properly in the mata command, which I have been unable to do.

. local t `"This is a "multi word token" within a string"'

. macro list _t
_t:             This is a "multi word token" within a string

The expected result would be

This,is,a,"multi word token",within,a,string

if I understand his wishes correctly.

But beyond that, I think he's making the case, with which I agree, that any such capability should be a naturally discoverable element of Stata's macro commands.

[Bjarte Aagnes]

ref 306 on compound quotes...

local mylist  `"" a " " b " "a b" "   " `"""""' ",,," "' 

mata: st_local("name1", ustrregexra(ustrregexrf(invtokens(char(96)+char(34) :+ (tokens(st_local("mylist"))) :+ char(34)+char(39) + ","),",$",""),"', `", "',`")) 
  
mac li _name1

tokenize `"`name1'"' , parse(",") 

forvalues i = 1(2)11 {
    
    di "element " %02.0f `i' `" betwen bars |``i''|"'
}

 mac li _name1
_name1:         `" a "',`" b "',`"a b"',`" "',`"""""',`",,,"'

. 
. tokenize `"`name1'"' , parse(",") 

. 
. forvalues i = 1(2)11 {
  2.     
.     di "element " %02.0f `i' `" betwen bars |``i''|"'
  3. }
element 01 betwen bars | a |
element 03 betwen bars | b |
element 05 betwen bars |a b|
element 07 betwen bars |   |
element 09 betwen bars |"""|
element 11 betwen bars |,,,|

[David Fisher]

Support for pooling proportions in the meta suite (e.g. incorporation of Marc Arbyn's metaprop), please

With a little work to calculate your own effect size and standard error (e.g., the logit of the proportion and the standard error), you can feed this into the meta suite. The same applies if you had a single group mean, or whatever else.

Furthermore, note that the user-written metan package (currently v4.05 29nov2021) now natively handles proportions, among much else.

[Ben Jarvis]

Essentially frames should behave more-or-less as files do. My biggest frustration with frames is that they are not nearly as flexible as files are. This is very basic data processing stuff that can be done easily in other programming environments, but is a real headache with Stata. For example:

• appending frames together should be easy
• merging/crossing/joining frames should be easy. frlink is just not very good.
• should be easy to add batches of observations to an existing frame, not simply doing it one at a time, like frame post does.

I think this has been mentioned elsewhere in this forum, but I want to reiterate those points.

[William Lisowski]

Based on the discussion today at

https://www.statalist.org/forums/for...nual-with-loop

I would like to see the output of help limits explain that these two lines

   # of lines in a program                    3,500              3,500
   # of bytes in a program                  135,600            135,600

also apply to the number of lines and bytes between open and close braces, e.g. within a foreach loop.

[Joro Kolev]

Stata regression commands that employ transformed variables (e.g., -xtreg-, -areg-, etc.) should offer the option to output/save somewhere the transformed versions of the variables.

Sergio Correia, whose -reghdfe- by now is probably more widely used than Stata's native -xtreg- and -areg-, would do a great service if his -reghdfe- offers the option to output the transformed variables.

[Jesse Wursten]

Stata regression commands that employ transformed variables (e.g., -xtreg-, -areg-, etc.) should offer the option to output/save somewhere the transformed versions of the variables.

Sergio Correia, whose -reghdfe- by now is probably more widely used than Stata's native -xtreg- and -areg-, would do a great service if his -reghdfe- offers the option to output the transformed variables.

I think you can do this with -hdfe-

[Zumin Shi]

I wish etable and collect functions provide a choice of journal style, similar to the Endnote software's output style. Currently, it is time consuming to write the code to generate the regression table I want.

[Fahad Mirza]

Not sure if this is a useful suggestion, but would be really nice if Stata allows users to save multiple tabs within the Do File. Not only this, but then be able to refer those tabs while executing a dofile.

[Yingyi Lin]

getting standardized coefficients from svy: regress (similar as what we had in the regress, beta options)