[#6133] Improvement(core): Supports get Fileset schema location in the AuthorizationUtils

[Abyss-lord]

What changes were proposed in this pull request?

Supports get Fileset schema location in the AuthorizationUtils

A Fileset can be uniquely identified using metalake.catalog.schema.fileset. The logic for retrieving the schema is as follows:

Check the type of catalogObj.type():

• If it is RELATIONAL, determine whether the provider is Hive. If it is a Hive table, retrieve its LOCATION property.
• If it is FILESET, determine whether it implements HasPropertyMetadata:
  1. If it does, use the schemaPropertiesMetadata() method to retrieve the path.
  2. If it does not implement HasPropertyMetadata, check whether it contains any Fileset objects:
     1. If it does not, convert the catalog object to FilesetCatalog and retrieve its LOCATION property.
     2. If it does contain Filesets, retrieve all the Fileset instances and add their respective Fileset paths.

Why are the changes needed?

Fix: #6133

Does this PR introduce any user-facing change?

No

How was this patch tested?

local test.

[Abyss-lord]

Hi @xunliu , could you please review this PR when you have time? I’d really appreciate your feedback.

[xunliu]

@Abyss-lord
I think we needs to consider to add an integration test, It's include Schema is Fileset.

[jerqi]

Why do we need to judge HasPropertyMetdatadata?

[Abyss-lord]

@jerqi Is the fileset catalog guaranteed to implement the HasPropertyMetadata interface?

[jerqi]

Some of the fileset catalogs have the property location although they implement the HasPropertyMetdata interface.

[Abyss-lord]

The current processing logic is

1. If the provider is Hadoop, use HasPropertyMetadata to get the location of the Schema.
2. In other cases, consider adding paths to all filesets below the Schema.
   WDYT, @jerqi

[jerqi]

The 2 seems weird for me. Is this consistent with Hive schema implement.