[Bert Lloyd]

The ability to add custom metadata to .gph files and retrieve using graph describe, similar to graph describe's current ability to retrieve r(dtafile), r(command), etc.


(Some long-ago discussion here: https://www.statalist.org/forums/for...generated-pdfs)

[Ghasem Yadegarfar]

I wish to see mainly all predict options for different survival functions such as Cox-PH, strreg, stpm2cr, weibul,..For example, to be able to see

1. C-statistics (Harrell’s C index) (Concordance C-index) with 95% CI
2. area Under the Curve (AUC) with 95% CI
3. linear prediction, HR,..

[Clyde Schechter]

The C-statistic is already built-in after -stcox- with the -estat concordance- command. For any other model, you can get it from Roger Newson's -somersd- package, available from the Stata Journal. Similarly you can get the linear prediction and hazard ratio with the -predict- command following all survival analysis commands (except those that are estimated only in accelerated failure time metric)--you just need to specify the -xb- or -hr- options respectively..

[Ghasem Yadegarfar]

Hi Clyde. Thank you so much for your very kind help and explanation. somersd is not running after stfpmcr function. Please can you explain how to get C-stat after running xb predict option? Many thanks

[William Lisowski]

Perhaps this discussion should be continued at the original post of this problem, at

https://www.statalist.org/forums/for...redict-options

[William Lisowski]

Having just learned of do, nostop I wish that were included in the dropdown menu on the toolbar of the Do-file Editor so that it could be used with temporary do-files created by the Do-file Editor.

[Brian Poi]

I would like to see support for time subscripts with variable names when the data have been -tsset-, similar to syntaxes supported by Eviews and Fame. Oftentimes I need the value of a variable at one specific time, perhaps to use as a base year. Then instead of typing

. tsset mdate, monthly
. ...
. summ x if mdate == tm(2021m1), mean
. scalar base = r(mean)

I could simply type

. tsset mdate, monthly
...
. scalar base = x[2021m1]

or something similar. You'd get an error message if the data are not currently -tsset-. Ostensibly, you could extend this to -xtset- data with double subscripts, though I would be perfectly happy with just support for -tsset- data.

[William Lisowski]

That is an interesting idea. It might be preferable to implement this as a tsobs() function which returns the observation number you seek; I have wanted this in other contexts. Your example would become

scalar base = x[tsobs(2021m1)]

where tsobs() would use the tsset values to infer that 2021m1 is a tm()-style date.

In the meanwhile, if your time series has no gaps, and the delta is 1, you can calculate the observation number directly, as below.

. tsset mdate

Time variable: mdate, 2000m1 to 2021m12
        Delta: 1 month

. scalar base = x[tm(2021m1)-mdate[1]+1]

. display base
-.24886636

. list in `=tm(2021m1)-mdate[1]+1', clean

        mdate           x  
253.   2021m1   -.2488664

[Ameer Muhammad]

https://www.statalist.org/forums/for...tween-4-groups

Wasn't expecting to see Stata 17 come out today, but it looks like it has a lot of good new stuff.

Anyway, here's what I was going to add to my wishlist for Stata 18--I assume this hasn't been implemented in 17 (which I've ordered, but don't have my download yet.): let's have the ability to open a separate browser for each frame.

thank you, yeah stata 17 is good, python integration is real change.
please can you answer my this post requesting you ? https://www.statalist.org/forums/for...tween-4-groups

[William Lisowski]

I would like to see the Data Editor "Paste Special" dialog support more of the options that the import delimited command supports. In particular, allowing the user to choose whether the variable names should have their case preserved, rather than unavoidably set to lower case, would make life on Statalist a lot easier when members post data as a copy/pasted table to accompany example code that expects CapitalizedVariableNames.

[Kim Lyngby Mikkelsen]

In Stata 18 I would like "putexcel" to control colum width and perhaps row hight in Excel
As it is now, with putexcel I can format nice tables in Excel, but I still have to open the spreadsheet in Excel e.g, to autofit the colum width!

I imagine the Stata codes could look like this:

putexcel set ...

putexcel C:G, columnwidth(autofit) /* to autofit the width of the entire colums C to G */
putexcel C:G, columnwidth(14) /* to set the width of colum C to G to 14 */
putexcel 2:4, Rowheight(autofit) /* to autofit the hight of row 2 to 4 */
putexcel 2:4, Rowheight(35) /* to set the hight of row 2 to 4 to 35 */

[Maarten Buis]

That sounds more like what the new table command was made for.

[Christopher Bratt]

Allow for integration with RStudio (specifically with R Markdown), similar to the current integration with Jupyter Notebook.

Current alternatives to integrate R and Stata from RStudio include the R packages Statamarkdown and RStata.

[Ken Simons]

Function words(s, m, n) to return words # m through n in string s. Integers m and n may be negative, to count from the end of the string.

Also parts(s, m, n, parsingChars) to return parts # m through n in string s, using the parsing characters (like for tokenize) in string parsingChars. Again, m and n may be negative to count from the end. If parsingChars is a space, this is the same as words(s, m, n).

I've found such functions repeatedly helpful when working with strings.

[George Ford]

I may have posted this on the 17 thread, but I (and many others who have asked "how to") would like a command that conducts a resampling of a single series, group of series, and possibly an entire matrix.

resample varlist, noreplace block() generate() stub() number()

default is with replacement (override with noreplace, which is like shufflevar)
block() permits a block bootstrap
generate() or stub() permits a new series contained the resample series or group
number() permits fewer or more than the original sample (using . for missing or extending data for >N).