
We copied our Dataverse production environment to a sandbox with the configuration only option. We then loaded partial data back into the sandbox. We loaded 11 million records into one of the tables and the Dataverse administration page indicates that the table is consuming 40 GB of space. That's almost 4K per record of consumption, but the data itself on average is closer 500 bytes per record. What likely happened is the configuration-only copy brought across all indices from the production environment, even though most of the indices are likely not needed in the sandbox. What's needed by customers is 1) a way to see where a table's consumption is actually going and 2) a way to view and drop indices that are not needed. Without this level of control and visibility, Dataverse administrators and those writing the checks for a company cannot predict their Dataverse consumption costs and, thus, cannot budget for the use of the platform.
You can connect on premise SSMS if you like. You should be able to view the dB like an on premise dB server.
I used this link
https://learn.microsoft.com/en-us/power-apps/developer/data-platform/dataverse-sql-query
Please accept this as a solution if it helps, thank you