Counting the number of variables with same stub

[John Larsson]

Hello,
I am trying to count the number of variables that use the same stub and verify that there are at least two such variables.
I want to do something like:
ds stub*
I know this will then return the the number of variables so named in r(varlist).
Can anyone tell me how I can then check if there are at least two elements in the list?
Or perhaps there is a better way of doing this?
Regards,
John L.

[Robert Picard]

Here's two ways, the first is preferred if you do not have Stata 13

sysuse auto, clear

ds m*

local nwords :  word count `r(varlist)'

display `nwords'

display wordcount("`r(varlist)'")

[Roberto Ferrer]

-wordcount()- is also available in Stata 12.1. So I wonder if the statement

Here's two ways, the first is preferred if you do not have Stata 13

was made based on the belief that it's not available before Stata 13, or rather there is some other, underlying reason.

[Robert Picard]

Indeed, I wasn't very clear. Before Stata 13, the limit on Stata strings was 244 characters. So using any string function to manipulate a macro could yield unexpected results. Here's a quick example in Stata 9:

. forvalues i = 1/30 {
  2.         local vlist `vlist' a_rather_long_variable_name`i'
  3. }

. 
. local nwords :  word count `vlist'

. 
. display `nwords'
30

. 
. display wordcount("`vlist'")
9

The same code works well in Stata 13:

. forvalues i = 1/30 {
  2.         local vlist `vlist' a_rather_long_variable_name`i'
  3. }

. 
. local nwords :  word count `vlist'

. 
. display `nwords'
30

. 
. display wordcount("`vlist'")
30

So if you are using Stata 12 or earlier, you should always use the macro extended functions (see help extended_fcn) to manipulate macros.

[Roberto Ferrer]

Excellent Robert.

Thanks for the details!

[Marvin Aliaga]

I think I have a similar questions. I receive data in excel spreadsheets from different community based organizations. My job is to verify that that the data they submit are clean. One of the things I need to do is to quickly count the number of variables. If the datasets have fewer or more variable than expected I will have to let them know. There are supposed to have 30 variable in their datasets. Is there a way to quickly count variables and if they are != 30 variables? Thank you
Marvin

[Nick Cox]

The number of variables in the dataset is built in as c(k)

. sysuse auto, clear
(1978 Automobile Data)

. di c(k)
12

[Marvin Aliaga]

Thank you Nick!

[Dan Daugaard]

Is it possible to use the number of stubs we have counted as the limit in the forvalues command? For example, using the stub counting method presented in #2 above

sysuse auto, clear
ds m*
local nwords : word count `r(varlist)'
display `nwords'

followed by a forvalues command which uses the stub count - this one isn't correct but intended to present the idea

forvalues i = 1/`nwords' {
 }

Thank you, Dan

[William Lisowski]

This question is also posted, with a reply, at the following topic.

https://www.statalist.org/forums/for...ch-observation