[Clyde Schechter]

A frequent question seen on this Forum arises when a string variable is included in a regression command and the result is a "no observations" error message. While, in a sense, this is true, it would be more informative to the user if the error message in this situation were "string variables not allowed" or something to that effect, preferably accompanied by a list of the offending variable names.

[wbuchanan]

alejoforero knowing if it was compressed previously would also require rewriting any command that results in creating a new variable, since the new variable may not be storage optimized. It might not be a ton of overhead, but it would add a ton of work for the developers for something that may be easier for end users to ultimately manage.

[alejoforero]

The same way that Stata is able to know if data has been sorted before trying to sort it again, it should be able to know if data has been recently compressed before trying to compress it again. This should speed things up for large data for which compress is particularly useful.

[Niels Henrik Bruun]

When I have several Stata versions running, a little version number, e.g. 2, occurs at the top left just after the Stata icon.
It would be very nice if the do-file editor and the data editor had the corresponding version number shown at the top left just as their original Stata version has.

[Stephen Weinberg]

I love the new table features, but I would love to be able to sort the rows by a particular value (for example, in descending order by the contents of a given column). I am frankly astonished that I cannot find any sorting features. (I have stumbled across the user-written myaxis ado file, thank you Nicholas Cox.)

I wish graph bar/graph hbar allowed me to add values at the end of the bars based on other variables, or let me change the opacity or color of the bars based on some other variable. (For example, I want my reader to be able to see which bars are computed from low sample sizes.)

[Jared Greathouse]

FernandoRios boy wouldn't that be perfect.

[FernandoRios]

I would like to see more "Stata" command have their "Mata" equivalent. Perhaps this would be too much for complex commands, but for simpler ones, that require fewer checkups, it would be convenient.
(for example a regress, logit, probit, etc)

[Hendri Adriaens]

re unifying excel support across stata and mata, actually unify all commands and syntax, eg. import delim imports variable names lowercase by default, you don't need to declare the first row to contain variable names. import excel preserves casing of variable names by default, and needs instructions to read the variable names from the first row.

[Hendri Adriaens]

- speed up import of data
- don't hang the ui when a longer process is busy (like importing data)
- treat missings as missings, so x > 50 doesn't result in true for missings, but in missings (this requires the use of proper data types without missings defined as large numbers)
- support debugging, especially of for-loops
- support formatting excel files within stata (so unify features with mata) and introduce open and close commands to avoid reading&writing on each command (is slow, see above)
- introduce a proper project management which allows for refactoring variable names, file names, etc across files in your project

[Emil Alnor]

Network Analytics (Econometrics and Visualization)

I will second on this. Exponential family random graph models and tools for visualising network data is currently the only reason I have to R sometimes. The whole familiy of tools for doing SNA seems missing from Stata. I know there are some user written commands, but it would be convinient if the StataCorp could sort them through and add some to "base" Stata.

[Jaakko Markkanen]

My personal wishlist is mostly focused on making Stata faster:

1. Implement the estimation of otherwise identical regression equations with only different outcome variables by using only one command (like in fixest for R). Looping different outcome variables currently requires unnecessary matrix calculations.
2. Allow import delimited and export delimited to read/write compressed files.
3. Native support for parquet files.
4. frlink should support 1:m
5. Append for frames
6. Collapse to frame
7. Allow frlink and merge based on label values. Currently, it's not possible to encode and label string IDs from different data sources before merging because the keys won't match. This is very inconvenient with large data sets.
8. Related to the previous part, a new data type 'category' would probably be a good idea. Or a new, more efficient data type for strings (like in data.table/datatable).

Furthermore, I would love to see the Stata native adoption of gtools, reghdfe and boottest

[FernandoRios]

Maarten Buis and Jean-Michel Galarneau
While not official, it is possible to use margins to work with other functional forms including splines if you have -f_able- installed.
Note: I just realized there is a typo in my f_rcpline ado, but the example should still work (update has been submitted), and you can get the file from the attachment)

sysuse uslifeexp, clear
f_rcspline yr=year, nk(5)
reg le year yr*
f_able yr?, auto
margins, dydx(year)
margins, at(year=(1900(5)2000)) plot
margins, dydx(year) at(year=(1900(5)2000)) plot

The problem, however, would be adding the "jump" since my approach only works well with continuous variables. Not discreet. and in this case, Spanish flue would be a function of the year, which i do not include here.

also, until the package update is done, you can get the -f_rcpline.ado- file attached.
Best wishes

[Maarten Buis]

Margins and marginsplot support for restricted cubic splines.

I agree. However, you can make margins and marginsplot work if you only want predictions. If you want to use margins to compute marginal effects, then you are out of luck right now, and it would indeed be nice if that changed in future versions of Stata.

// some example data
sysuse uslifeexp, clear

// to spice things up we do a cubic spline and discrete jump
mkspline ys = year , cubic
gen spanish_flu = year == 1918

reg le ys* i.spanish_flu

// we can use margins for predictions
margins, over(year)
marginsplot,                            ///
    recastci(rarea) ciopts(astyle(ci))  ///
    plotopts(msymbol(i)) scheme(s1mono)

[Leonardo Guizzetti]

Find easy way to flag (e.g. or color a command line) to fix parentheses not balanced error message. Not sure if this is solved in the past.

The do file editor has the option to highlight matching braces, so just move the caret to one brace and the corresponding one will be highlighted.

[Million Tadesse]

Find easy way to flag (e.g. or color a command line) to fix parentheses not balanced error message. Not sure if this is solved in the past.