[Feature] Specify table API which Rest catalog needs support

[jerry-024]

Search before asking

• I searched in the issues and found nothing similar.

Motivation

Specify table API which Rest catalog needs support.

Solution

I have searched other catalog APIs like iceberg Rest catalog and unity catalog,

• the APIs we need support with are
  • getTable
  • listTables
  • dropTable
  • createTable
  • renameTable
  • alterTable
  • getTableLocation： as it only use when clone table
• need discuss
  • tableExists: as we needn't use this API in common so getTable is enough

Other catalog APIs

iceberg

• listTables
• createTable
• tableExists
• dropTable
• renameTable
• loadTable
• invalidateTable
• registerTable

unity catalog

• listTables
• getTable
• deleteTable
• tableExists

Are you willing to submit a PR?

• I'm willing to submit a PR!

[adrian-wang]

How about getTables to batch get tables to avoid network communication overhead?

[jerry-024]

What's the different between listTables and getTables?

[sundapeng]

Are there any plans to support the snapshot, branch, or tag API ?

[JingsongLi]

Are there any plans to support the snapshot, branch, or tag API ?

Good point, these need some refactor, we will discuss them in follow issues.

[JingsongLi]

How about getTables to batch get tables to avoid network communication overhead?

@adrian-wang Do you know which scenarios need this batch API? I cannot see it in computation engines.

[sundapeng]

Are there any plans to support the snapshot, branch, or tag API ?

Good point, these need some refactor, we will discuss them in follow issues.

I'm looking forward to the discussion.

[sundapeng]

How about getTables to batch get tables to avoid network communication overhead?

@adrian-wang Do you know which scenarios need this batch API? I cannot see it in computation engines.

@JingsongLi , APIs for tables can optimize certain management scenarios, such as listTables typically returning table names and getTables needing to return full table objects. Multi-table join scenarios may also benefit.

Based on my previous experience, batch APIs for partitions can also significantly improve performance in some queries.

Any thought? @adrian-wang

[adrian-wang]

@sundapeng +1