Issues with dtable command.

[Tanner Smida]

I recently updated to Stata18 and was hoping to try out the dtable command. When using nhanes data, the command works as advertised.

However, when I attempt to generate a new string variable or new byte variable from the nhanes data - in this case, I chose to generate another Sex variable [generate byte Sex1 = sex] - I get an r(9999); error message. Basically, any new variable I generate causes this error to appear. Any advice on how to rectify this? Thank you for your time.

[Clyde Schechter]

I am unable to reproduce your problem on my setup. The same code runs fine:

. webuse nhanes2l, clear
(Second National Health and Nutrition Examination Survey)

.
. dtable age weight bpsystol i.sex i.race

----------------------------------------
                             Summary    
----------------------------------------
N                                 10,351
Age (years)              47.580 (17.215)
Weight (kg)              71.898 (15.356)
Systolic blood pressure 130.882 (23.333)
Sex                                    
  Male                     4,915 (47.5%)
  Female                   5,436 (52.5%)
Race                                    
  White                    9,065 (87.6%)
  Black                    1,086 (10.5%)
  Other                       200 (1.9%)
----------------------------------------

.
. gen byte sex1 = sex

. tab sex1 sex

           |          Sex
      sex1 |      Male     Female |     Total
-----------+----------------------+----------
         1 |     4,915          0 |     4,915
         2 |         0      5,436 |     5,436
-----------+----------------------+----------
     Total |     4,915      5,436 |    10,351

.

Moreover, there isn't supposed to be a 9999 error code. So I think something is likely wrong with your Stata installation. If I were you, I would uninstall Stata 18 and reinstall it from scratch. If the problem persists, I think you should contact Technical Support about this.

[Tanner Smida]

Thank you, Mr. Schechter. I should clarify - the error message appears when I try to rerun the dtable command with the new variable. When attempting to use any string variable, I also get the 'string vars not allowed in varlist' error code before the r(9999); code appears.

[Tanner Smida]

For others that may be having the same problem - this error appears when the varname has a capital letter.

[Leonardo Guizzetti]

Can confirm - if the first letter is capitalized then -dtable- gives error cod 9999.

webuse nhanes2l, clear
rename age Age
dtable Age

[Clyde Schechter]

Ah, I misunderstood the original post. I, too, now have difficulties with -dtable- when there is a variable beginning with an upper case letter.

. clear*

. webuse nhanes2l
(Second National Health and Nutrition Examination Survey)

. dtable sex

-----------------
       Summary  
-----------------
N          10,351
Sex 1.525 (0.499)
-----------------

. rename sex Sex

. dtable Sex
r(9999);

Then, it occurred to me that perhaps this was a problem because, for a factor-variable, -dtable- expects an i. prefix. Perhaps it's just an incorrect error message that is supposed to instead be a warning about factor variable notation. So I tried that and got, to my surprise:

. dtable i.Sex
factor-variable operators not allowed on latent variable Sex
r(198);

This kind of message ought to be coming out of -sem-, not -dtable-. And -dtable- should work with all legal variable names.

So, yes, something is definitely wrong with -dtable- and variable names beginning with upper case.

While the StataCorp team does follow the Forum, it wouldn't hurt to report this directly to Tech Support.

[Leonardo Guizzetti]

Also note the use of a continuous variable in my earlier example (age), so definitely not limited to factor variables.

[Jeff Pitblado (StataCorp)]

dtable has a parsing bug that is causing it to exit with unhelpful error messages when a variable name with a leading capital letter is specified in the varlist.
This also happens in options continuous() and factor().
This should be fixed in the next Stata 18 update.

[Andrea Correa]

Hello. I am using dtable for summary statistics for the variables specified but separately for each category using, by (var)

I receive a table containing summary statistics that display column percentages. This means that the percentages on each column add up to one hundred.

I need a table that shows row percentages. Do you know how can I get that? Thanks!

This is the code:

. dtable usdpercapita i.zona i.sexo , by(die10_c, nototals test)

this is the table I get:

----------------------------------------------------------------------------------------------------------------
die10_c. ¿con qué frecuencia revisa y toma en cuenta los sellos de advertencia
SIEMPRE CASI SIEMPRE ALGUNAS VECES RARA VEZ NUNCA Test
----------------------------------------------------------------------------------------------------------------
N 1,661 (30.1%) 566 (10.3%) 532 (9.6%) 494 (8.9%) 2,267 (41.1%)
usdpercapita 404.456 (802.428) 349.860 (416.370) 360.858 (507.553) 302.784 (320.349) 272.651 (297.909) <0.001
Área geográfica
URBANA 1,469 (88.4%) 501 (88.5%) 446 (83.8%) 409 (82.8%) 1,820 (80.3%) <0.001
RURAL 192 (11.6%) 65 (11.5%) 86 (16.2%) 85 (17.2%) 447 (19.7%)
Sexo
HOMBRE 524 (31.5%) 199 (35.2%) 175 (32.9%) 166 (33.6%) 955 (42.1%) <0.001
MUJER 1,137 (68.5%) 367 (64.8%) 357 (67.1%) 328 (66.4%) 1,312 (57.9%)
----------------------------------------------------------------------------------------------------------------

As you can see, the percentages are by columns, not by rows.

[shem shen]

how can i use dtable in Stata 16?

[Leonardo Guizzetti]

how can i use dtable in Stata 16?

the short answer is that you can’t. dtable was released in Stata 18. To use it you would need to upgrade.

[Brian Nguyen]

I have a silly question: if you are using dtable for the first time, do you have to create a word doc and then send the table to that designated word doc? i'm a little confused how dtable works. I am able to create the table fine, but i just can't export it. Every time I run: export myfile.docx
I end up getting export error: unknown subcommand "myfile.docx"

I just assumed dtable would create a file automatically on my desktop or in my stata folder.

I know this is perhaps a stupid question. any ideas?

[Rich Goldstein]

one of the options to the command is "as()" where you put in parens the fileformat you want (e.g., "docx") - note that there are docx options; see

h dtable

and search for, or scroll down to, the "as()" option where you can click on "docx options" to see what is available

[Brian Nguyen]

Do I need to specify a location where stata should create the doc, or is that something that is automatically to desktop?

[Brian Nguyen]

Jeff Pitblado (StataCorp)