[EpiData-list] error in CHK file: cannot apply opGTE to missing operands

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Fri Sep 16 01:46:52 CDT 2011


Many thanks for this quick response..!

Ruwan

On 9/15/11, epidata-list at lists.umanitoba.ca
<epidata-list at lists.umanitoba.ca> wrote:
> Allow for missing data and put this first in the logic. Then the comparison
> should be OK. What I show is that you allow either MOUS or MOUSTR to be
> missing and then the comparison is not necessary.
>
> CONSISTENCYBLOCK
>    CHECK "impossible # treated nets"   ((mous=.) or (moustr=.) or (MOUS >=
> MOUSTR))
>    CHECK "impossible # children sleeping under nets"    ((enfants=.) or
> (dormi=.) or (ENFANTS >= DORMI)
> END
>
> If you do not allow only one of the fields to be missing, you have to run
> the following first and fix all the records that do not comply:
>
> consistencyblock
>   check "Inappropriate missing values for MOUS or MOUSTR" ((mous<>.) and
> (moustr<>.))
> end
>
> This is easy to do in Analysis, but you still have to be careful about
> missing values, because you cannot trust the result of < or > on a missing
> value.
> In Analysis, of course, your intention is to select bad records, so the
> logic is reversed.
>
> read yourdata
> select (mous<>.) and (moustr<>.) and (MOUS<MOUSTR)
> list id mous moustr
> select
> select (enfants<>.) and (dormi<>.) and (enfants<dormi)
> list id enfants dormi
> select
>
> Jamie
>
> On 2011-09-15, at 5:54 AM, epidata-list at lists.umanitoba.ca wrote:
>
>> I receive the following error with the CHK file coding that follows:
>>
>> CHECK MOUS >= MOUSTR
>> cannot apply opGTE to missing operands
>>
>> And here's the coding:
>>
>> CONSISTENCYBLOCK
>>    CHECK "impossible # treated nets"    MOUS >= MOUSTR
>>    CHECK "impossible # children sleeping under nets"    ENFANTS >= DORMI
>> END
>>
>
> _______________________________________________
> EpiData-list mailing list
> EpiData-list at lists.umanitoba.ca
> http://lists.umanitoba.ca/mailman/listinfo/epidata-list
>


More information about the EpiData-list mailing list