[Feature][cdc] Extract schema change from record

### Search before asking

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

### Motivation

Currently, in cdc synchronization, there may be a situation where the schema comes later but the data comes first, which may lead to a dead wait on the write side.

If we extract the schema change from the data, it is in the realm of invincibility and can never go wrong.

### Solution

_No response_

### Anything else?

_No response_

### Are you willing to submit a PR?

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

The issue names no files, tests, or entry points. Start by tracing the CDC synchronization write path and how schema and data events are handled when data arrives first. Done means the schema change is extracted from the data flow and the dead-wait scenario is prevented, with tests covering the ordering case.

Written by the indexing model from the issue text.