[AURON #1739] Support LIMIT with OFFSET

[yew1eb]

Which issue does this PR close?

Closes #1739

Rationale for this change

What changes are included in this PR?

Are there any user-facing changes?

How was this patch tested?

[yew1eb]

cc @richox

[Copilot]

Pull request overview

This PR adds support for SQL LIMIT with OFFSET functionality to the Auron query engine, implementing the feature for Spark 3.4 and later versions.

Key Changes

• Extended limit operations to support offset parameter, allowing queries like SELECT * FROM table LIMIT 5 OFFSET 2
• Changed parameter types from Long to Int for consistency with Spark's internal representations
• Updated protobuf schema to include offset field with reduced integer size (uint64 → uint32)

Reviewed changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
NativeTakeOrderedBase.scala	Added offset parameter; updated executeCollect to handle offset by dropping rows after sorted merge
NativeLocalLimitBase.scala	Changed limit parameter type from Long to Int
NativeGlobalLimitBase.scala	Added offset parameter and updated native execution to pass offset to protobuf
NativeCollectLimitBase.scala	Added offset parameter; updated executeCollect and native execution to handle offset
Shims.scala	Added getLimitAndOffset methods for Spark 3.4+ exec nodes; updated factory method signatures
AuronConverters.scala	Updated converters to extract and pass limit/offset pairs from Spark exec nodes
AuronExecSuite.scala	Added comprehensive tests for limit with offset across different exec types
NativeTakeOrderedExec.scala	Updated case class to include offset parameter
NativePartialTakeOrderedExec.scala	Changed limit parameter type from Long to Int
NativeLocalLimitExec.scala	Changed limit parameter type from Long to Int
NativeGlobalLimitExec.scala	Added offset parameter to case class
NativeCollectLimitExec.scala	Added offset parameter to case class
ShimsImpl.scala	Implemented getLimitAndOffset methods for Spark 3.4/3.5; added effectiveLimit helper
limit_exec.rs	Added skip field; implemented execute_limit_with_skip function; updated display and statistics
from_proto.rs	Updated deserialization to handle offset in limit and sort operations
auron.proto	Changed FetchLimit and LimitExecNode from uint64 to uint32; added offset field

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[richox]

this nested execution plan will break metric nodes tree. i suggest to add a new offset field to SortExec and handle offset logics in SortExec::execute()