[Feature] Schema Change can be generated by data in CDC ingestion

### Search before asking

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

### Motivation

Especially in Kafka, there may be disorder, and the Schema Change may arrive later than the data. If there is a checkpoint difference between them, it may cause data blockage.

For schema changes, they can actually be generated from data and do not necessarily rely on cdc.

### Solution

_No response_

### Anything else?

_No response_

### Are you willing to submit a PR?

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

No files, tests, or entry points are identified in the issue. Start by locating the CDC ingestion handling for schema changes and the behavior around out-of-order Kafka data; done should include tests showing that schema changes can be generated from data without blocking ingestion.

Written by the indexing model from the issue text.