Lakehouse do not see data from last few days (in SQL endpoint nor Lakehouse view)

[PiotrAntoniszyn]

Hi lads,

We've encountered an issue as we do not see new data in our lakehouse from last few days (both in SQL endpoint view and lakehouse view). We thought it might be sync delays between Lakehouse and SQL Analytics Endpoint, because we saw in files that those data were properly added. Also we saw those data in another lakehouse where we have shortcut to the first one's tables.

We tried to refresh SQL analytics endpoint Metadata using REST API, but scripts did not see any table what was really weird. Additionally, when checking tables' properties in SQL Endpoint view there wasn't any date in Last synced updates from OneLake.

Somehow this case was resolved, after few days, we created another shortcut from broken lakehouse and suddenly all the data there were available through SQL endpoint again, also we've seen date in Last synced updates from OneLake.

My question is, can someone explain to me what was the real issue of that and why the shortcut creation resolved this issue? Is there some kind of refresh underneath the shortcut creation process?

[MJParikh · Accepted solution]

According to my analysis, this behavior points to a temporary metadata desynchronization between the Lakehouse and the SQL Analytics endpoint. The key factor is that the SQL endpoint metadata cache sometimes doesn’t refresh automatically when new Parquet files land in OneLake, especially after schema changes or delayed commit operations. The Lakehouse view shows the data because it reads directly from the file system, while the SQL endpoint relies on cached metadata.

When you created a new shortcut, Fabric triggered an internal metadata rebuild process. This refreshed the SQL endpoint’s mapping to the underlying OneLake tables, forcing a rescan of the files and re-registering the table schema. That’s why your “Last synced updates from OneLake” timestamp appeared again and the missing data became visible.

To prevent this in the future:

1. Manually refresh metadata
   In SQL Analytics endpoint, run
   ALTER TABLE <table_name> REFRESH METADATA;
   This forces a resync between OneLake and the endpoint.

2. Use REST API or PowerShell
   Call the Lakehouse metadata refresh endpoint periodically for automation.
   Endpoint:
   POST https://api.fabric.microsoft.com/v1/workspaces/{workspaceId}/lakehouses/{lakehouseId}/refreshmetadat...

3. Avoid stale schemas
   After schema evolution (column add, type change), always trigger a metadata refresh before querying via SQL endpoint.

4. Monitor “Last synced updates from OneLake”
   If it stops updating for multiple tables, it’s a sign the metadata service didn’t complete its sync cycle.

Creating a new shortcut worked because it implicitly forced Fabric to rebuild table metadata lineage, similar to a full refresh.

[v-menakakota]

Hi @PiotrAntoniszyn ,
Thanks for reaching out to the Microsoft fabric community forum. 

I would also take a moment to thank @MJParikh   , for actively participating in the community forum and for the solutions you’ve been sharing in the community forum. Your contributions make a real difference.

I hope the above details help you fix the issue. If you still have any questions or need more help, feel free to reach out. We’re always here to support you 

Best Regards, 
Community Support Team 

[v-menakakota]

Hi @PiotrAntoniszyn ,

I hope the above details help you fix the issue. If you still have any questions or need more help, feel free to reach out. We’re always here to support you .

Best Regards, 
Community Support Team