Re: st: checking data using assert

[Nick Cox <njcoxstata@gmail.com>]

This feeds my prejudice that -count- is often overlooked. See

SJ-7-1  pr0029  . . . . . . . . . . . . . . .  Speaking Stata: Making it count
        . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
        Q1/07   SJ 7(1):117--130                                 (no commands)
        discusses count used with a loop over observations
        or variables

http://www.stata-journal.com/sjpdf.html?articlenum=pr0029 for freely
available .pdf

I would make the actions conditional on finding problems (positive
count). There is no need to use -assert- or create a new variable.

qui count if dod <  dob

if r(N) > 0 {
            di _new as err  "date of death before birth date':"
            list id dob dod  if dod < dob, noobs N
}

Nick
njcoxstata@gmail.com


On 29 May 2013 03:00, Raoul Reulen <r.c.reulen@bham.ac.uk> wrote:

> I am currently running several do-files in order to check data. I would like Stata to check whether a certain statement is true, e.g. date of birth (dob) < date of death (dod). If it is not true I want it to display a message saying what the problem is and display the ID numbers for those records were the statement was not true. Lastly, I want to list several variables (e,g. id, dob, dox) to inspect the actual problem.
>
> I have done the following, but wondered if there are easier/more efficient ways of doing it?
>
> ******************************
> .gen byte tag = dod <  dob
> .qui:  levelsof   id    if tag==1
> .cap: assert  tag==0
> .if _rc di _new as err  "date of death before birth date; see id:   " as text "`r(levels)'"
> .list id dob dod  if tag==1 , noobs N
> .drop tag
> ******************************

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/