You should split up the actions into two different coordinator workflows. This
is because your workflow will start as soon as the coordinator's data/time
dependency is met, and in this case, your second action has a different
dependency than the first. After you separate out the second action into its
own workflow, use a new coordinator with the data dependency set to those 7-day
aggregates, and it will get triggered only when they are ready i.e. the 1st
workflow (the daily aggregator) has finished its job.
Mona
On Friday, May 8, 2015 10:14 AM, Narayanan K <[email protected]>
wrote:
Hi
Just wanted to check if some one has any suggestion ..
Narayan
On Thu, May 7, 2015 at 9:52 PM, Narayanan K <[email protected]> wrote:
> Hi
>
> I have a workflow which has 2 actions run by a coordinator running daily.
>
> The first one is daily aggregation - reads today's data, aggregates
> and stores into today's daily folder.
>
> The second action needs 7 day's daily aggregated data including
> today's generated by the first action i.e it reads today's daily
> aggregated data generated by the first action + the aggregated daily
> data generated by the first action in the last 6 runs of the
> coordinator.
>
> Eg : If today is 20150510, first action generates
> /user/abc/daily_interm/20150510.
> Second action needs 7 days as input :
> /user/abc/daily_interm/20150504,/user/abc/daily_interm/20150505,/user/abc/daily_interm/20150506,/user/abc/daily_interm/201505047,/user/abc/daily_interm/20150508,/user/abc/daily_interm/20150509,/user/abc/daily_interm/20150510
>
> How can I form the input for the second action from the coordinator to
> the workflow ?
>
> Thanks
> Narayan
