No. That is legal code, but it doesn't do what you want.
missing(x1-x6) is true if the difference between the two variables x1
and x6, i.e. the expression x1-x6, is missing, not if any of the
variables x1 through x6 is missing.
There are at least two ways to see this:
1. Experiment with made-up data.
2. Look at the help. You want -missing()- to support
-missing(<varlist>)- but it doesn't. Its arguments must be separated by
commas. (Conversely, a <varlist> cannot contain commas. If it did then
commands based on <cmdname> <varlist> , <options> could not work
reliably.)
Nick
[email protected]
Ashim Kapoor
would missing(x1-x6) work? I have seen that sometimes.
On Thu, Nov 13, 2008 at 6:34 PM, Nick Cox <[email protected]> wrote:
> Note that what you must type, or otherwise get Stata to see, is
>
> missing(x1,x2,x3,x4,x5,x6)
>
> That is, Stata will not replace an ellipsis ... with what you intend
or
> imply.
>
> That said, you can go
>
> list id if missing(x1,x2,x3,x4,x5,x6)
>
> The new variable is not essential for this purpose.
Johannes Geyer
> gen miss0 = missing(x1,x2,x3,...,x6)
> list id if miss0 == 1
Sham
>> I wish to list the ID numbers of those individuals in my dataset
>> that have at least one missing value out any six variables (x1-x6).
>> At the moment I have used the following logic:
>>
>> list responder_status ID if x1==. | x2==. | ///
>> x3==. | x4==. | x5==. | x6==.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
