Jason, You described perfectly what I am looking for. The "Running Sum" operator is what would be useful for solving my problem.
Edwin On Thu, Dec 12, 2013 at 7:36 AM, Jason Pickering < [email protected]> wrote: > Hi Knut, > This is not quite the same thing. There may be data for time periods > earlier than 2013 of course. > > But as I have implemented this requirement other places in the form of > reports/scripts to generate this data, if the following data is reported > > Dec 2012 9 > Jan 2013 10 > Feb 2013 15 > March 2013 14 > .. > > The cumulative "indicator" for this would be > Dec 2012 9 > Jan 2013 19 > Feb 2013 34 > March 2013 48 > ... > > These sorts of indicators are needed when calculating things like "Number > of people who have ever started ART", where "New ART starts" are reported > each month. So instead of having the facility to calculate "Number of > people who have every started ART", it can be calculated with a > cumulative/running total instead from the new figures reported each month. > I think this arises from the lack of a "RUNNING SUM" operator for the time > dimension in DHIS2, which probably should be there. > > Best regards, > Jason > > > > On Thu, Dec 12, 2013 at 2:53 AM, Knut Staring <[email protected]> wrote: > >> I think you could just take the yearly data for 2013. This should add up >> what is available so far this year. >> >> Sent from my mobile >> On 12 Dec 2013 01:55, "Jason Pickering" <[email protected]> >> wrote: >> >>> Hi Eddie, >>> >>> You certainly could, but there is no way to do this directly with DHIS2 >>> I think . You could use windowing functions of Postgresql [0] to calculate >>> the cumulative sum, or create your own function with a custom function with >>> PL/R which has been described in the documentation outlined here [1]. There >>> are various ways to do this in R, either with the base "cumsum" function or >>> with more advanced methods in the "timeSeries" package depending on your >>> exact needs. >>> >>> The simplest method however might be to use Excel [2] . >>> >>> However at the moment, there is no way to do this directly with DHIS2 as >>> far as I know. But if others know more direct methods, it would be good to >>> know! >>> >>> Regards, >>> Jason >>> >>> >>> [0] >>> http://www.postgresonline.com/journal/archives/47-How-to-calculate-Running-Totals-and-Sums-in-SQL.html >>> [1] http://www.dhis2.org/doc/snapshot/en/user/html/apcs06.html >>> [2]http://support.microsoft.com/kb/214149 >>> >>> >>> On Wed, Dec 11, 2013 at 1:50 PM, Edwin Mulwa <[email protected]> wrote: >>> >>>> Hi, >>>> >>>> I would like to create a report that has a field(s) which is a >>>> cumulative total calculation involving multiple data elements. For example, >>>> I have data elements A and B and I am doing monthly cumulative reporting >>>> such that the value for element A in Feb will be the value of A in Jan + >>>> the new value for B in Feb, the value for A in March will be the value of A >>>> in Feb + the new value for B in March, etc. >>>> >>>> Can we create such a report in DHIS? >>>> >>>> Edwin >>>> >>>> _______________________________________________ >>>> Mailing list: https://launchpad.net/~dhis2-users >>>> Post to : [email protected] >>>> Unsubscribe : https://launchpad.net/~dhis2-users >>>> More help : https://help.launchpad.net/ListHelp >>>> >>>> >>> >>> _______________________________________________ >>> Mailing list: https://launchpad.net/~dhis2-users >>> Post to : [email protected] >>> Unsubscribe : https://launchpad.net/~dhis2-users >>> More help : https://help.launchpad.net/ListHelp >>> >>> >
_______________________________________________ Mailing list: https://launchpad.net/~dhis2-users Post to : [email protected] Unsubscribe : https://launchpad.net/~dhis2-users More help : https://help.launchpad.net/ListHelp
