Hi everyone,
My team was running an older version of Datahub for a while, but have just recently upgraded to the latest version and have enjoyed exploring the new search UI. We had a couple quick questions about some of the things we’re seeing though.
For entities with no browse path beyond the platform name (i.e. an empty BrowsePathV2 aspect), the drop down triangle is still there but clicking it doesn’t display any new hops (see picture 1). Obviously because there are no browse paths, there is nothing to load here, but I think it could be a bit confusing to users if the dropdown action still happens but nothing loads. It may not be clear that they have to click the platform name itself. I noticed on the demo site, all entities with an empty browse path are placed into a “Default” folder (see picture 2). I thought that seemed like a good solution and was wondering how that was done. Is the “Default” folder a manual addition, some sort of transformer, an upcoming feature, or some other fourth thing?
The second question was about the order filters are listed on the search page. What’s the criteria for how these are listed? It’s not the order they were selected in, but rather seems to follow the pattern of Entity type > container > environment > platform. Is there a reason for this choice?
Thanks!
first, we auto-generate aspects for entities that we ingest and that’s where we generate a default browse path which will try to use info from the entity urn and if there’s no good info to discern a default browsePathsV2 aspect, we fall back to “Default” as you’ve called out. I’m not sure how you would have gotten these entities in without at least a default path generated… how did you ingest those entities? and are they datasets?
second part - do you mean the order that the filter dropdowns appear in? or the order in which we show you the selected filters underneath them? (the little tags with an “x” on it)
if it’s the first case, for the order of the filter dropdowns, there’s an order that we’ve defined based on what we thought was the best UX/highest priority filters. However, there are dynamic filters that we get back from the search endpoint as well and we just display those in the order we receive them
Hey, thanks for the reply Chris. For the “Default” browse path, the entities in question include datasets, charts, & dashboards ingested via a combination of recipes and custom built connectors. I did a little more looking and it seems like the entities in question actually don’t have a browsePathsV2 aspect at all and that’s why they’re not showing up. I’m not quite sure what happened there but we should be able to just rerun those ingestion pipelines to fix things. It’s good to know that that’s the default behavior though so we can replicate that in our custom connectors as we make changes to those.
For the second part, I was referring to the selected filter tags underneath the main filter (shown in picture). They don’t seem to appear in the order they’re selected (either via the filter drop downs or by navigating the browse path tree), because sometimes new filter tags are placed before existing ones. The order appears to be consistent too, so it is definitely being enforced in some way. Like in the provide picture, the filter tags are always container > environment > platform. I was just curious if this was a conscious decision or not. If it was, I’d think it’d make more sense to go environment > platform > container.
but yeah great questions