arrow

GH-41818: [C++][Parquet] normalize dictionary encoding to use RLE_DICTIONARY (#41819)



### Rationale for this change

There're some points:

1. https://github.com/apache/arrow/blob/main/cpp/src/parquet/encoding.cc#L444-L445 . encoding is not passed in Encoder
2. But, it's RLE in decoder: https://github.com/apache/arrow/blob/main/cpp/src/parquet/encoding.cc#L1607 it will be detect and normalized in other place, like:
3. https://github.com/apache/arrow/blob/main/cpp/src/parquet/column_reader.cc#L876

We'd better unifying them

### What changes are included in this PR?

Unify dict encoding to `RLE_DICTIONARY`.

### Are these changes tested?

No need

### Are there any user-facing changes?

No

* GitHub Issue: #41818

Authored-by: mwish <maplewish117@gmail.com>
Signed-off-by: Antoine Pitrou <antoine@python.org>