dhis2-devs team mailing list archive

[Lars Helge Øverland <larshelge@xxxxxxxxx>]

Hi Jason,

the issue here is that "dis-aggregation" of data elements is not allowed.
So for a "monthly" data element, weekly or daily data is ignored (data with
higher frequency).

The frequency/period type of a data element is derived from the period type
of the data sets which it is assigned to. When a data element is member of
data sets with different period types, the behavior is undefined, ie. the
"first" data set is just selected randomly. If there are no data sets, a
yearly period type is assumed. We could of course change this behavior.

regards,

Lars



On Sat, May 11, 2013 at 2:15 PM, Jason Pickering <
jason.p.pickering@xxxxxxxxx> wrote:

> OK, so let me get this straight. Lets say I have a data element which
> belongs to a weekly and to a monthly dataset. Some orgunits report it
> weekly, others monthly. This is what is happening here in Zambia.
>
> Which one wins? The weekly or the monthly?
>
> It would seem that we would need to aggregate all of the data which was
> reported with a weekly period, to weekly, and to monthly, and all of the
> data which was reported monthly to monthly?
>
> I will need to confirm whether or not this data element is part of another
> dataset.
>
> What happens if the data element is not part of a data set at all?
>
> Regards,
> Jason
>
>
>
> On Sat, May 11, 2013 at 9:17 AM, Ola Hodne Titlestad <olati@xxxxxxxxxx>wrote:
>
>> Just to clarify:
>> You'll get many aggregated period types in the datamart of course, but
>> only one period type per data element can be used when looking/searching
>> for raw data values to put into aggregation calculations during datamart
>> export.
>>
>> So your monthly data values might be ignored in the datamart export
>> process if quarterly data values exist for the same data elements. And we
>> use the dataset period type to know what data values to look for. At least
>> that's how I've understood this.
>>
>> Lars, correct me if this is wrong.
>>
>> Ola
>> ------
>> On 11 May 2013 09:10, "Ola Hodne Titlestad" <olati@xxxxxxxxxx> wrote:
>>
>>> Hi Jason,
>>>
>>> Are these data elements (with missing monthly vslues in datamart) also
>>> part of another dataset with e.g. quarterly collection frequency (period
>>> type)?
>>>
>>> Only one period type per data element is supported in datamart and this
>>> is derived from the dataset's period type. So if the data element is member
>>> of two data sets with different period types you will only get one of these
>>> period types in the datamart.
>>>
>>> Ola
>>> ------
>>> On 9 May 2013 18:02, "Jason Pickering" <jason.p.pickering@xxxxxxxxx>
>>> wrote:
>>>
>>>> Hi devs,
>>>> I am having some issues with some data elements not appearing in the
>>>> data visualizer (and other modules which depend on datamart). Using the
>>>> latest 2.11 file.
>>>>
>>>> I get expected results from the analytics pivot tables, but nothing for
>>>> the "Past twelve months" as a relative period in the DV.
>>>>
>>>> Looked at the aggregateddata value table where this should appear, and
>>>> see a some periods, but no monthly values (this is how the data is
>>>> collected, by month).
>>>>
>>>> SELECT startdate from period where periodid in (SELECT DISTINCT
>>>> periodid from aggregateddatavalue where dataelementid = 80);
>>>>  startdate  |  enddate
>>>> ------------+------------
>>>>  2012-01-01 | 2012-12-31
>>>>  2012-01-01 | 2012-03-31
>>>>  2012-01-01 | 2012-06-30
>>>>  2012-04-01 | 2012-06-30
>>>>  2012-07-01 | 2012-09-30
>>>>  2012-07-01 | 2012-12-31
>>>>  2012-10-01 | 2012-12-31
>>>>  2013-01-01 | 2013-06-30
>>>>  2013-01-01 | 2013-03-31
>>>>  2013-01-01 | 2013-12-31
>>>> (10 rows)
>>>>
>>>> Looking at the data values table, just to be sure.
>>>>
>>>> SELECT startdate,enddate from period where periodid in (SELECT DISTINCT
>>>> periodid from datavalue where dataelementid = 80) ORDER BY startdate DESC;
>>>>
>>>> startdate  |  enddate
>>>> ------------+------------
>>>>  2013-03-01 | 2013-03-31
>>>>  2013-02-01 | 2013-02-28
>>>>  2013-01-01 | 2013-01-31
>>>>  2012-12-01 | 2012-12-31
>>>>  2012-11-01 | 2012-11-30
>>>>  2012-10-01 | 2012-10-31
>>>>  2012-09-01 | 2012-09-30
>>>>  2012-08-01 | 2012-08-31
>>>>  2012-07-01 | 2012-07-31
>>>>  2012-06-01 | 2012-06-30
>>>>  2012-05-01 | 2012-05-31
>>>>  2012-04-01 | 2012-04-30
>>>>  2012-03-01 | 2012-03-31
>>>>  2012-02-01 | 2012-02-29
>>>>  2012-01-01 | 2012-01-31
>>>>  2011-12-01 | 2011-12-31
>>>>  2011-11-01 | 2011-11-30
>>>> :
>>>> So, the raw data appears to be there.
>>>>
>>>> I see no obvious errors in the logs. I did check the "Monthly" option
>>>> for the data mart.
>>>>
>>>> I find this really strange. Any ideas what may be going on? For other
>>>> data elements collected in the same form, I do get monthly values.
>>>>
>>>> Regards,
>>>> Jason
>>>>
>>>>
>>>> _______________________________________________
>>>> Mailing list: https://launchpad.net/~dhis2-devs
>>>> Post to     : dhis2-devs@xxxxxxxxxxxxxxxxxxx
>>>> Unsubscribe : https://launchpad.net/~dhis2-devs
>>>> More help   : https://help.launchpad.net/ListHelp
>>>>
>>>>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~dhis2-devs
> Post to     : dhis2-devs@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~dhis2-devs
> More help   : https://help.launchpad.net/ListHelp
>
>