[Feature] Optimize batch multiple partitions inserting

### Search before asking

- [X] I searched in the [issues](https://github.com/apache/incubator-paimon/issues) and found nothing similar.

### Motivation

By default, batch sink should sort the input by partition and sequence_field to avoid generating a large number of small files. Too many small files cause poor performance, especially object storage.

We can not implement `SupportsPartitioning.requiresPartitionGrouping`. we need sequence.field to sort, otherwise we can't confirm what the last record is.

### Solution

_No response_

### Anything else?

_No response_

### Are you willing to submit a PR?

- [ ] I'm willing to submit a PR!

Start by locating the batch sink path and reading the SupportsPartitioning.requiresPartitionGrouping contract, then trace how sequence.field is handled. Define the expected partition and sequence ordering behavior and identify focused batch-sink tests to verify it; the issue names no files or tests, so the implementation entry point must be located first.

Written by the indexing model from the issue text.