Hi All,
I have the following requirement:
Suppose a process is as Start->A->B->C->D->E->End. Now, I would like to view the Avg Throughtput Time for Activity E.
There are cases where the process flows from A->E, B-> E, C->E and D->E. (different variations)
So the logic should then calculate the Average of the Time Differences between Activity E and the activities which occur just before Activity E. (could be any one of A / B / C / D).
We dont have the Endtime field. (it wouldve then been just a diff b/w Starttime - Endtime)
I hope the requirement seems clear. Your help would be highly appreciated.
Thank You.
Page 1 / 1
Hi @Keshav,
As far as I understand your issue correctly, you are looking for a PQL statement that measures a Subprocess Cycle Time.
DAYS_BETWEEN(
PU_FIRST(Case Table, Activity Table.EVENTTIME, Activity Table.Activity Column = Your Activity START ),
PU_LAST(Case Table, Activity Table.EVENTTIME, Activity Table.Activity Column = Your Activity END)
)
If you need further support, please do not hesitate to contact us again!
Best,
Justin
As far as I understand your issue correctly, you are looking for a PQL statement that measures a Subprocess Cycle Time.
DAYS_BETWEEN(
PU_FIRST(Case Table, Activity Table.EVENTTIME, Activity Table.Activity Column = Your Activity START ),
PU_LAST(Case Table, Activity Table.EVENTTIME, Activity Table.Activity Column = Your Activity END)
)
If you need further support, please do not hesitate to contact us again!
Best,
Justin
Hi @j.stegmaier
Thanks for the response but Im not looking for a particular Subprocess Cycle Time.
In the query shared by you, I would need to input 2 Activities: Start Activity and End Activity.
However in my scenario, I only wish to input End Activity. Any activity that comes directly before that should be considered as Start Activity and I need the average of the throughput times of all such cases. For instance, (refer attached screenshot) I need to find the Avg Throughput Time for Book Invoice. So I need my query to give the result as:
Avg(Time between Create Purchase Order to Book Invoice AND Change Price to Book Invoice AND Receive Goods to Book Invoice AND ) There could be different start Activities directly followed by Book Invoice. I need the Avg time considering all those scenarios.
Hope this gives some clarity n the requirement.
Thanks
Thanks for the response but Im not looking for a particular Subprocess Cycle Time.
In the query shared by you, I would need to input 2 Activities: Start Activity and End Activity.
However in my scenario, I only wish to input End Activity. Any activity that comes directly before that should be considered as Start Activity and I need the average of the throughput times of all such cases. For instance, (refer attached screenshot) I need to find the Avg Throughput Time for Book Invoice. So I need my query to give the result as:
Avg(Time between Create Purchase Order to Book Invoice AND Change Price to Book Invoice AND Receive Goods to Book Invoice AND ) There could be different start Activities directly followed by Book Invoice. I need the Avg time considering all those scenarios.
Hope this gives some clarity n the requirement.
Thanks
Hi @Keshav
I think you just need the a case when - statement:
AVG(
CASE WHEN
ACTIVITY_LEAD ( _CEL_P2P_ACTIVITIES.ACTIVITY_EN ) = Book invoice
THEN
DAYS_BETWEEN("_CEL_P2P_ACTIVITIES".EVENTTIME , ACTIVITY_LEAD ( _CEL_P2P_ACTIVITIES.EVENTTIME ))
END
)
Hope it helps.
BR
I think you just need the a case when - statement:
AVG(
CASE WHEN
ACTIVITY_LEAD ( _CEL_P2P_ACTIVITIES.ACTIVITY_EN ) = Book invoice
THEN
DAYS_BETWEEN("_CEL_P2P_ACTIVITIES".EVENTTIME , ACTIVITY_LEAD ( _CEL_P2P_ACTIVITIES.EVENTTIME ))
END
)
Hope it helps.
BR
Hi @Keshav,
what you also could try is using Source and Target.
E.g. apply the following in an OLAP to make it more visible:
Dimension 1: SOURCE(_CEL_P2P_ACTIVITIES.ACTIVITY_EN)
Dimension 2: TARGET ( _CEL_P2P_ACTIVITIES.ACTIVITY_EN)
Apply a component filter to select on flows where the second activity is always E: FILTER _CEL_P2P_ACTIVITIES.ACTIVITY_EN = E
KPI: DAYS_BETWEEN (SOURCE("_CEL_P2P_ACTIVITIES".EVENTTIME), TARGET("_CEL_P2P_ACTIVITIES".EVENTTIME))
Let me know, how it turns out.
BR
Matthias
what you also could try is using Source and Target.
E.g. apply the following in an OLAP to make it more visible:
Dimension 1: SOURCE(_CEL_P2P_ACTIVITIES.ACTIVITY_EN)
Dimension 2: TARGET ( _CEL_P2P_ACTIVITIES.ACTIVITY_EN)
Apply a component filter to select on flows where the second activity is always E: FILTER _CEL_P2P_ACTIVITIES.ACTIVITY_EN = E
KPI: DAYS_BETWEEN (SOURCE("_CEL_P2P_ACTIVITIES".EVENTTIME), TARGET("_CEL_P2P_ACTIVITIES".EVENTTIME))
Let me know, how it turns out.
BR
Matthias
Reply
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.