[Feature] Introduce a memory option to control column batch size

### Search before asking

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

### Motivation

Now it is hard to use `read.batch-size、orc.write.batch-size`, because every row has different memory size.
We can introduce a memory option to control column batch size

### Solution

_No response_

### Anything else?

_No response_

### Are you willing to submit a PR?

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

Start by tracing the read.batch-size and orc.write.batch-size options to understand how column batches are currently sized. Define the memory option and its behavior for rows with different memory usage, then add coverage for the resulting batch-size control. The issue provides no files, tests, or proposed solution, so the intended behavior needs clarification first.

Written by the indexing model from the issue text.