`long window` supports base table with existing data

**Describe the feature you'd like**

If data is loaded into the table before the deployment with long window option,
we should build the pre-aggregation table on the fly (asynchronous).

**Additional context**

Add any other context or screenshots about the feature request here.

The issue mentions 'long window' and building a pre-aggregation table asynchronously for existing data. Research the OpenMLDB codebase for long window handling and pre-aggregation logic. Look for existing table deployment and data loading paths. Determine how to trigger an asynchronous build process and integrate it with the existing data flow.

Escrito por el modelo de indexación a partir del texto del issue.