[Sebastian Kripfganz]

There is something that I find increasingly annoying with the Do-File-Editor in Stata 19. When I want to search for a word or something else in a do-file, it is impossible to see which occurence is currently hightlighted when I move through the do-file with "Find next". This is because by default all occurences of the same word are highlighted. The only solution that works for me is to turn of highlighting for matching words and the current selection in the preferences. It would be better if other matches were highlighted differently to the current selection. (But maybe I should just make another appointment with my optician.)

[Fahad Mirza]

GitHub/Git Integration will be very useful. Many folks (at least in my circle) work on impact evaluations and create reproducible package submissions to DIME. However, when collaborating in a team, it gets a bit difficult if multiple users use the same do file and that ends up creating conflicts which is tedious to resolve.

[Fahad Mirza]

Can we have the box plot also marks the mean? Not really a Stata 20 request but more like a general request.

[Radion Svynarenko]

Let us change the column width in Data Editor using syntax

[Chen Samulsion]

Let us change the column width in Data Editor using syntax

see -recol- which dynamically resize columns in Stata's data browser

recol from http://fmwww.bc.edu/RePEc/bocode/r
    'RECOL': module to Dynamically Resize Columns in Stata's Data Browser /
    recol examines your variable names and the content of your / variables in
    order to determine the appropriate display widths / for your columns in
    the data browser. It works with string, / numeric, and date variables.  /

[Leonardo Guizzetti]

Let us change the column width in Data Editor using syntax

This can be done already. If you go into the data browser/editor and try to manually change the width of a column, a warning pops up informing the user that column widths are stored as data characteristics and therefore will mark the dataset as changed. Programmatically, you can create a characteristics by the name of "_de_col_width_" for each variable you wish to change, whose value is the column character width.

For example, the below piece of code will make variable make unreadable in the auto dataset because the column width is now too narrow.

sysuse auto
char make[_de_col_width_] 10
browse

[John Mullahy]

A small request: A "match case" option for "Find" and "Find and Replace" in the Do-File Editor.

I'm running StataNow/SE 18.5 so it's possible this has already been implemented in newer versions.

[Clyde Schechter]

Maybe it's a flavor or a Windows vs Mac thing? I have such a match case option in version 19.5MP for Windows, and also back in version 17MP. And I have the impression that it dates back even farther than that, though I no longer have any other versions active on my setup, so I'm not certain about this.

[John Mullahy]

Thanks Clyde. It's quite possible I've missed something, but I've checked the Do-FIle Window Windows Preferences under Settings as well as the Edit Menu when a Do-File is active and haven't found it.

While life will somehow manage to go on without such an option, it would have made some coding this afternoon a lot easier had it been available.

[Leonardo Guizzetti]

A small request: A "match case" option for "Find" and "Find and Replace" in the Do-File Editor.

I'm running StataNow/SE 18.5 so it's possible this has already been implemented in newer versions.

I have this option on Windows using Stata 18, and I'm positive it existed in at least Stata 17, corroborating Clyde's response. From the Do file Editor, go to the Edit menu and select either Find or Replace and it's an option from both functions.

[John Mullahy]

I should have noted in #82 that I'm running StataNow/SE 18.5 on a Mac.

[Chinh Nguyen (StataCorp)]

There is something that I find increasingly annoying with the Do-File-Editor in Stata 19. When I want to search for a word or something else in a do-file, it is impossible to see which occurence is currently hightlighted when I move through the do-file with "Find next". This is because by default all occurences of the same word are highlighted. The only solution that works for me is to turn of highlighting for matching words and the current selection in the preferences. It would be better if other matches were highlighted differently to the current selection. (But maybe I should just make another appointment with my optician.)

I believe we use the system color for selections as the default color for selections in the Do-file Editor on Windows and Mac. In your case (and mine), that default color is very similar to the default color for highlighting matching selections in the Do-file Editor. You can simply change either the default selection color in Do-file Editor's Preferences > Color > Display Items > General > Selection color or the Do-file Editor's Preferences > Color > Display Items > Matches > Selection color. We'll have to look into coming up with better highlight colors in the future but it's always going to be possible for it to be very similar to someone's selection color.

[Chinh Nguyen (StataCorp)]

I should have noted in #82 that I'm running StataNow/SE 18.5 on a Mac.

In the Find bar that's used by the Do-file Editor (and other windows like Results and Data Editor) on the Mac, there's a little triangle next to the magnifying glass icon that you can click on to set options for finding text including enable case-sensitive searches as well as partial or whole words. It's admittedly a little subtle but that interface element was consistent with what other applications were using at the time it was implemented (which was quite a long time ago). The issue is that we don't have a lot of room on the Find bar to display all the possible options which is why we chose to use a menu (unless of course you make the window much bigger). What looks like empty space that could be used for displaying options is actually where we print out how many occurrences there are of a found item.

FYI, with a StataNow license you are eligible for a free upgrade to StataNow 19.

[Robert Grant]

Related to a post I'm writing in the general forum, I should log here the idea that it would be relatively easy for Stata to add Hamiltonian monte Carlo (e.g. the NUTS algorithm) as a sampler in bayes: and bayesmh. There are implementations (Stan, PyMC) that are open source, have been widely used for many years, and that have effective diagnostics and warnings metrics. These algorithms give a huge benefit in speed and stability for correlated posterior distributions.

[John Mullahy]

This request concerns specification of base() values in twoway area plots. As it stands one can specify a scalar value in a twoway area plot to define the area to be shaded. E.g. if I want to shade the area below a (faux) ROC curve down to the x-axis I could do this:

twoway ///
function y=ibeta(1,2,x), ra(0 1) recast(area) base(0)

It's unlikely that I would want to produce such a graph but it's realistic that I might want to produce a graph where the shaded area is that between the function and the 45-degree line. There is even a way to do this using the nodropbase option

tw ///
(function y=ibeta(1,2,x), ra(0 1) recast(area) nodropbase ) ///
(function y=x, ra(0 1))

My v20 Wishlist request is to provide a more general way to specify the base as a function of x. The nodropbase option is fine for what it does but the ability to specify a more general base for area shading would be valuable, e.g. something along the lines of

tw ///
function y=<some function of x>, ra(`L' `U') recast(area) base(<some other function of x>)