Shortcut is possible in Lakehouse (and KQL) but NOT in Warehouse. So if a DIM table is required in mulitple Warehouses then we need to load it to each Warehouse. What is the point of not having shortcuts in Warehouse?
Around the 5 minute mark, you showed an example of pulling a shortcut into another workspace, but it didn’t work because you “need to do some PySpark stuff to get it up and running” What PySpark stuff was needed?
Hi Adam, Patrick - thanks for the video. One question - Can we enable Direct Lake and Direct Query on the Shortcuts when we access the data in power bi? If Direct Lake fails, can we have a fallback plan to fetch the data using DQ mode from PBI to shortcuts?
Say you have a bunch of csv files in your datalake in Azure storage. If you create tables from them and then update the csv , then the tables don't seem to change. Also, what if you have "versions" of csv files? my_data1.csv, my_data2.csv .. can you read all csv files of the same structure ? You can do this with serverless sql by just using openrowset with an asterisk * as a wildcard. I find this more difficult in fabric.
Question - 2 users - user 1 has created shortcut (of a table) on workspace1, lakehouse1 (user 1 has permision on Azure storage). User 2 do not have permission on Azure storage (from where user 1 created shortcut). But user 1 gave permission to user 2 on his lakehouse 1 - would user 2 be able to create shortcut of this table to his lakehouse (let call it lakehouse 2) on different workspace - of course the tenant is the same?
Would these shortcuts be valid to access data outside of your own organization without really replicating the data? Like reading the data from another organization without really copying the data.
Referencing another shortcut, what will happen to the second shortcut when you drop the first? Will it still be working or will it break. From a data lineage perspective also very interesting 😁
Would love to get thoughts on if/when shortcuts could become a solution to bring together data from multiple SAAS solutions from other vendors. If I have understood the feature then this could unlock a lot of currently 'siloed' data sets. Will try and ask this on a live stream when get the chance :)
Azure, which is already a Microsoft product, and Amazon, the only truly external source, aren't sufficient. More options like Google Cloud Storage are needed.