Identify variables with a particular value

[Dan Frank]

Hello, I have a cross-sectional dataset with many variables, and I am attempting to compile a list (disaggregated by my groupvar, county) of instances in which any variable takes on the value of 5.

I have attempted to do this using the foreach command to loop through all variables, but I'm having some trouble.

When I use the code below, I get the "type mismatch r(109)" error message.

foreach var of varlist all {
by county: list if `var' == 5
}

Assuming this occurred because I was not restricting my loop to numeric variables, I attempted to isolate all of the numeric variables with

ds, has(type numeric)

However, when I apply the code above, it returns a list of all variables.

Does anyone have any tips on the best way to compile of list of cases in which a variable takes on a particular value?

Many thanks!
Dan

[Nick Cox]

Your code and indeed your question are a little hard to follow.

In the first segment, did you really type all or was it _all? Either way, that code will indeed fail if it meets a string variable.

Your closing question is about identifying cases (Stata says observations) for which a variable takes on a particular value and the answer is like

list if myvar == 5

However, most of your question seems more general. findname (SJ) is an effect a superset of ds with (in my view) improved syntax.

This example may indicate what you are looking for.

sysuse auto, clear
findname, type(numeric)
egen any5 = anymatch(`r(varlist)'), value(5)
findname, any(@ == 5)
bysort foreign : list `r(varlist)' if any5

---------------------------------------------------------------------------
-> foreign = Domestic

     +--------------------------+
     | rep78   headroom   trunk |
     |--------------------------|
 20. |     5        2.0       8 |
 43. |     5        2.5      11 |
 46. |     2        5.0      16 |
     +--------------------------+

---------------------------------------------------------------------------
-> foreign = Foreign

     +--------------------------+
     | rep78   headroom   trunk |
     |--------------------------|
  1. |     5        3.0      15 |
  5. |     5        2.0       8 |
  9. |     5        3.0      10 |
 10. |     4        2.5       5 |
 14. |     5        2.5      11 |
     |--------------------------|
 15. |     5        2.5      14 |
 16. |     5        3.0       9 |
 17. |     5        2.0      11 |
 19. |     5        3.0      15 |
 22. |     5        2.5      14 |
     +--------------------------+

In short there are two techniques:

1. Finding the observations with any 5.

2. Finding the variables with any 5

which can naturally be combined.

You need to install findname before you can use it. Start with

search findname

and find the most recent download location.

Note that with its any() option findname just ignores type mismatches.