Hello, dear KNIME users 
I am again struggling with one task
I have data where I should take sum of data, but this aggregation should be summed up daily, weekly, monthly by each user
The output should be like this
My solution is to group them separately and after all grouping to concatenate them, But seems like it does not work at all
Can you help me please with this aggregation, please 
I have attached in my input file in here sum.xlsx (13.1 KB) mine.xlsx (22.7 KB)
Hello again @Karlygash !
The idea of how you are solving this sounds right, so Iâm guessing it is a small mistake in the config for one of the nodes.
Looking at your âmine.xlsxâ sheet, it appears to me that it isnât the aggregation that has gone wrong, but that somehow the whole data that is against your name has also been applied to âAinurâ, so is it maybe an issue with a join somewhere?
e.g.
Tricky to tell exactly how. Are you in a position to share the workflow? 
Good Day @takbb 
âmine.xlsxâ sheet - here I included only two employees and here only names are different, the rest of data are same. I have added one more employee, in order to show that the there are other employees and they also should be grouped by employee as well
Yes, I can share , but seems like my workflow kinda mess ) 
case.knwf (89.4 KB)
Hi, Donât worry I am not judging workflow, and besides yours is a lot tidier then mine are when Iâm still working on them.
One thing though, could you re-upload the sum.xlsx that the this workflow uses as it hasnât been included in the workflow folders. thanks
I think I can see a likely issue. Probably it shouldnât actually be concatenating those groupings, but should instead be joining them each individually to the main flow, using the grouping columns (Employee Name + week/month/day column) in the aggregations as the joining key. I believe that what was happening in yours was that the concatenated (grouped) data was being joined back but against just Employee name so was being multiplied up to a large number of âduplicatedâ rows.
I couldnât actually run your flow without the data file, but Iâve tried to work it through anyway. Attached is some logic for just that part, using the sum.xlsx from your initial post, so youâd have to play around with some of the config but hopefully it makes sense.
Period_aggregations.knwf (45.7 KB)
The input data was 73 rows, and so is the output, which hopefully is what you require. Onto the end of each of the rows it has appended Emp Hours for Month, Emp Hours for Week and Emp Hours for Day
The group by was done much the same way that you did. I chose to make keys out of Year+Week and Year+MonthNumber, only because it future proofs it if you ever had data for multiple years.
A question that you posed inside your workflow was how to turn a Date&Time Difference from 3H 15M format into 3.15.
This can be achieved by changing the config to return Minutes rather than Duration
I hope buried in the above are the answers you need, but obviously fire away with more questions if you have them.
(Though I should warn that I am meeting some friends in a pub in London this evening for the first time in over a year, so donât believe anything I tell you tomorrow! 
)
oh, sorry. I have just seen your reply
Thank you very much
I always forgot to uncheck reset workflow when I export it.
As usually perfect solution, I ll try to apply it today and upload worflow here
(oh, London 
miss this city so much 
)
Have a nice evening with friends
aha⊠I see you found a âDuration to Numberâ node So now I have learned something too!
case - takbb.knwf (183.3 KB)
Iâve upload a copy of your flow with some changes. One thing I noticed was that you were grouping days by Employee Name and Day of Month, but that will hit snags when you have more than one monthâs data, if that can happen, so I changed it to group by Employee Name and Apply Date.
I added an addition at the bottom that may or may not be useful to you. As things stand, what appears at the end is totals appended to every line. You might find it useful to have these totals displayed only on the final line of the day/week/month groupings for each employee,
I made a component for that a few months ago, and Iâve included it here as a demo.
What this box at the bottom does is create three âgroup keysâ using the same columns used for the earlier groupings. It then runs a component (three times⊠one for each grouping) that finds the first and last row in each of the groups. It then uses rule engine to clear the value for each row that isnât the last of the group.
That might be useful. If not just ignore it!
Iâm off now. So hope you have a good evening.
Hi, @takbb !
Thanks for your workflow, you really helped me
Youâre welcome @Karlygash , and thanks for ticking the solution! 
Looking forward to your next challenge, and Iâm sure @aworker is too!
Haha ha, with these so rewarding hearts by @Karlygash , definitely !
Cheers !
@takbb @aworker hello
Hope you are doing well
I have question related to this topic
In addition, for new row weekly total I used add empty rows, but seems like used it incorrectly
Hope you can help me 
Tried to use BIRT reporting, it was really new for me) Did not find proper settings yet
I have uploaded output sample and workflow without resetting =)
contingent worker_beta.knwf (814.9 KB) emp_data_out.xlsx (20.6 KB)
Thatâs OK @Karlygash. Iâm not at my pc this morning so cannot assist at the moment. Iâll catch up later if you are still needing assistance
@takbb , ok)) thanks)))
No problem at all @Karlygash. I like very much your posts and really enjoy helping you if possible. @takbb has contributed the most to this thread and I believe masters by now much more than me what is now in place. I will be of little help then 
@takbb will certainly reply soon !
Iâll definitely contribute if at a future step I can be of any help
All the best !
Ael
I have something, but it is mess))
Did not get the final result yet
Thank you so much!
I forgot to reset it 
contingent worker_beta.knwf (863.0 KB)
