Best Of
Re: YEARWEEK Function and Modes
Sadly this seems to be one of the functions that is not homologous to its MySQL counterpart, so unless someone has tried it, I don't think we'll have much extra documentation (You can always test and share your findings here, from my initial test any other number results in the same behavior as 11).
You could hack your way around this by using a DATE_ADD or DATE_SUB to move your actual date a couple of days so that the displaced calendar would align with weeks Sat-Fri, but this becomes tricky when you approach the beginning or end of the year.
YEARWEEK(DATE_ADD(`dt`,INTERVAL 1 DAY),11)
Will give you weeks starting on Saturday, but as explained if might work weirdly towards the start or end of a year. In here you can see that it considers 12/31/2022 as Week 202301.
Re: Data exploration beyond what users have access to
Do you want users to instantly gain access or should it go through a review? Should the request come through Buzz or via another option?
A quick and dirty option is to use the mailto protocol in an HTML hyperlink where you can specify to, subject and message body so you can send it to an automated email box:
<a href="mailto:email@example.com?cc=secondemail@example.com, anotheremail@example.com, &bcc=lastemail@example.com&subject=Mail from our Website&body=Some body text here">Send Email</a>
You can also use the CONCAT to dynamically include the dataset ID in the subject or body depending on your use case.
Re: How do I remove duplicate rows in a card drilldown?
Add another column that aggregates a field by using the count or sum, for example, and it will remove the duplicates. You can do this by dragging another field over and then click the pencil icon next to it and choose one of the aggregation options. Make sure you don't have anything that is in the sorting properties
Re: Custom User Profile Fields and be able to use them as Trusted Attributes for PDP
I know this thread is a bit old, but I'd like to add my thoughts on this feature. While it seems we can now define our own attributes which is really nice, it still requires those values come from IDP/SSO if you want those new attributes to be automatically inherited and locked down so the values on the profile cant be edited.
In my company, our PDP is very complex and we have a concept called data personas. Essentially, any given employee can where multiple 'hats' and may have 2 or more data personas. Each persona determines what datasets they have access too, as well as what data is visible. This allows us to give different filtering and object visibility rules to the same user depending on a given dataset. So his/her first data personas may allow them to see a 'Finance' dataset and have PDP filtering rules specifically tailored to that finance dataset. While the other persona allows them to see 'Operations' datasets with different PDP rules tailored to that persona.
So what would solve all of this in my world would be two things:
- Allow custom attributes to be table driven just like PDP automation that we can control to supplement any IDP provided attributes. In large companies, sometimes its not feasible or IT restricted where adding data to peoples SSO/ActiveDirecty/IDP profiles isn't allowed. That way we can put 'rules' in this table that contain the user, the name of a custom attribute, and the value we've set for them. Those then show up on their profiles and the trusted attributes screen and can be leveraged for PDP.
- Allow attributes to have sub attributes. This way, I can have a 'Data Persona' custom attribute, that also has sub attributes.
For example, lets say we have a Manager who has 2 personas. One called 'Ops Manager', one called Financial. It would be nice to see this kind of setup (everything in bold is a custom attribute, anything next to a minus sign would be a custom sub-attribute)
Data Persona
- Name | Finance
- Country | USA
- State | New York, Texas, California
- Name | Ops Manager
- Region | Southwest
This would allow me to greatly simplify PDP since I can infer their attributes directly. For every finance dataset, I can just add basic dynamic rules to my PDP table by putting a value like "domo.policy.managed_data_persona.name.finance" and PDP would automatically unfold the sub attributes to filter the data where the Country is USA and the States are New York, Texas, California. You could even have situations where you want BOTH personas taken into account for filtering. So you might do "domo.policy.managed_data_persona" and then all the sub-attributes are taken into account as OR rule evaluations so this user would see data where they see USA, New York, Texas, California OR where the Region is Southwest on the same dataset.
Where as right now, I have to generate individual rules/policies for each individual user for every dataset, every column and every value that they are allowed to see based on our internal persona data. It ends up being millions of rows to manage, and keeping up to date when their profiles change or they move to a different role/area of the business etc.
Apologies for the longwinded post. I just feel like PDP and Attributes could go much further and be much more flexible than they are today.
Thanks!
DomoDork
Re: Aggregate and ETL question
If you can have all of the records from your top dataset in your output dataset then you can do a left join on your dataset instead of an inner join. Then in your card do a count of the id file from your TEST dataset to get the total records that match
if you just want the entire count then I’d recommend running your TEST dataset through a group by tile and count in there and output it to a separate dataset