Hello, As part of the Datahub adoption in our ente...
# advice-metadata-modeling
c
Hello, As part of the Datahub adoption in our enterprise, we have a requirement to ingest and display a '_*business name*_' for the datasets and schema fields, so that our consumer is able to understand the business intent behind a given dataset/schema field. As we look into the current Datahub metadata model and documentation, we wanted to reach out to you with the below two options and seek your thoughts on what would be more appropriate or if you have any other solve that would meet our business requirement. Please review and advise. If this is not the right channel/thread for this question, please let me know as well. [ Tagging a few others from our team - @aloof-author-52810 @dry-zoo-35797 @nice-window-93693 @aloof-ram-72401 @creamy-postman-32779 @rhythmic-agent-54913 @swift-nail-32514 ] a. Utilize "*jsonProps*" at schema fields level to represent the business name of a given field. Per the definition in Datahub, "For schema fields that have other properties that are not modeled explicitly. Use this field to serialize those properties into a JSON string". Today, "jsonProps" is not searchable but we could make it searchable and make an upstream PR to Datahub. We would also need to make changes to the UI to display "jsonProps" in the schema tab of a given dataset against each field. Use "*customProperties*" at the Dataset level to represent the business name of the dataset. b. Extend metadata model by adding an additional "*business name*" to the schema field's definition and dataset's definition, so we could ingest and display it in the UI. We will have to make changes to the metadata models and UI and make an upstream PR to Datahub if this would provide value to other enterprises as well. c. Utilize glossaryTerms to represent the "*business name*" of a given schema field and dataset. d. Utilize Tags to represent the "*business name*" of a given schema field and dataset. e. Any other proposal as advised by Datahub.
thank you 1
m
I think what you described is what glossaryTerms are for.
m
@creamy-thailand-75908: we would advise using the Glossary terms to represent the business name for the fields or the dataset.
c
Hello @mammoth-bear-12532 Thanks for getting back to us on this requirement. We internally brainstormed this further. To add a little bit more to this need, it's more than just being able to add glossary term. The new business name field that we are talking about here equates to more an "*alias*" of a table column, which may not be human understandable always. So, business would like to add an additional "alias" to a given field. Same at the dataset level as well. Many of the dataset names from source systems are sort of "cryptic" to understand and having an "*alias*" other than the name and description would immensely help our business partners.
m
@green-football-43791 this is interesting... maybe the recent
label
extension you added to the field can be used for this
@creamy-thailand-75908 one challenge we had while thinking about it is... how would you prioritize whether to show the "description" of the field or the "business label"
c
@mammoth-bear-12532 - Label extension would certainly work for us at the field level. As we understand, the label extension is not being displayed in the UI, but perhaps the challenge is if there is not much real estate to display both description and the new "business label" field in the schema section without showing a horizontal scroll bar? a. Could we suggest adding a first line to the description field and include the "business label" if it's populated? b. Or could we suggest building a global setting that allows users to prioritize seeing description over "business label". We would be happy to partner and contribute through a PR. Thanks!
m
cc @green-football-43791
g
hey @creamy-thailand-75908! that would be great
I think (a) rendering the label on a new line if populated and leaving the table as is otherwise makes sense-- you can tag me in a PR!
s
Hey guys, any update on this point? I would like to see the field labels from looker in datahub but I cannot find them. Thanks