MCP should detect existing datasets before creating duplicates

What is the problem or goal you're trying to solve or accomplish?

When users create datasets via MCP, it may create a new dataset even if one already exists for the same table. This results in duplicate datasets — one created via UI (physical, lowercase schema) and one via MCP (virtual, uppercase schema) — pointing to the same underlying table. This causes split DAR permissions and confusion during dashboard export/import.

How are you solving it currently?

Users are advised to avoid using MCP for dataset creation and to use the UI or API instead. Existing duplicates need to be cleaned up manually by identifying which charts reference the old datasets and reassigning them.

What is your recommended solution?

Before creating a new dataset, the MCP should check if a dataset already exists for the same table in the workspace. If one is found, it should either reuse it or prompt the user to confirm whether to create a new one or use the existing dataset.