Delete/flag rows... a little bit advanced

Solved!
emate
Level 5
Delete/flag rows... a little bit advanced

Hi All,

I had this problem few days ago : https://community.dataiku.com/t5/Using-Dataiku-DSS/Delete-flag-rows-how-to-start/m-p/5199#M3356

And @Alex_Combessie  helped me a lot, but....  it turns out that my issue is more complex [attached new example], so I want to create a final flag to keep only these rows with all Product groups that my company has but within each country seperate. So for Country A I will keep rows with product group 1,2,3, but for country B I will keep only products from group 1 and 8 since "3" is not my company product group in this particural country.

My first idea is to use to create multiple columns  flags seperate for each country and then remove it with some conditions?


Thanks for any suggestions.

KR
Mateusz

 

0 Kudos
1 Solution
Alex_Combessie
Dataiker Alumni

A hint:

the Group recipe ๐Ÿ˜‰

.

.

.

The solution:

hint.gif

โ€ƒ

View solution in original post

7 Replies
emate
Level 5
Author

Attached file

0 Kudos
Alex_Combessie
Dataiker Alumni

Hi,

From your screenshot, I understand you want to keep only the rows of a product that is sold by your company, for each country (new requirement).

You can achieve this with a very similar flow as the one I shared in my previous example. The group recipe would have group keys by product group and country.

I have attached the new example project export.

Let me know if this fits your need.

Cheers,

Alex

0 Kudos
emate
Level 5
Author

Hi again!

Yes, I want to do the same as previously but I want to that within each Country separately.So for example, even though for Country = A & Company = MyC we have product group = 3 it is not flagged for country B, since for country B  and MyC there is no case.

I aprriciate all your help and I dont want to push it to much, but for now I'm working in training DSS enviroment and I have my training project where I can upload files and work on that and I dont think I am able to import the example you shared. If im wrong please correct me, if I can't could you please share a screen on how this has changed since previous version?

Thank you,
M

0 Kudos
Alex_Combessie
Dataiker Alumni

Hi,

Thanks for the precisions. It is possible your Dataiku admin has disabled project imports.

Do you see this "Import project" button?

Screenshot 2020-03-19 at 11.17.32.png

If not can you ask your Dataiku admin to import the project archive? I think that is the most efficient way compared to a screen recording.

Also, as you are currently training on DSS, I highly recommend this "From Excel to DSS" tutorial: https://academy.dataiku.com/latest/guide/analyst/excel-to-dss/index.html

Hope it helps,

Alex

emate
Level 5
Author

Hi again,

No, unfortunately I have limited access and I can't create/improt projects on my own. As I said I can only work within my "Training Project" created by an admin.

I will try and ask for this import,  maybe I will be lucky, but thank you anyway! ๐Ÿ™‚

M

0 Kudos
Alex_Combessie
Dataiker Alumni

A hint:

the Group recipe ๐Ÿ˜‰

.

.

.

The solution:

hint.gif

โ€ƒ

emate
Level 5
Author

Hi again,

Just one more question, since I'm not sure they will upload this project, I have to workaround that myself, maybe you will give me a hint. I tried to manipulate grouping recipe to get correct result. So my idea was to group it using 2 variables, Country and Product Group and than flag it. But it is working only partially. As you can see in attached screen it didnt flag Product Group=3 for Country B with "MyC" and that is correct (even though product group 3 is MyC for Country A).

But at the same time it flagged as "MyC" product group =8 in Country = A, but it shouldn't.

Thanks.

0 Kudos