[Bug] merge_update_columns: [] updates all columns

### Is this a new bug?

- [x] I believe this is a new bug
- [x] I have searched the existing issues, and I could not find an existing issue for this bug

### Which packages are affected?

- [ ] dbt-adapters
- [ ] dbt-tests-adapter
- [x] dbt-athena
- [x] dbt-athena-community
- [ ] dbt-bigquery
- [ ] dbt-postgres
- [ ] dbt-redshift
- [ ] dbt-snowflake
- [ ] dbt-spark

### Current Behavior

I have a model with
```
models:
- name: model_name
config:
materialized: incremental
incremental_strategy: merge
unique_key: key
merge_update_columns: []
table_type: iceberg
```

On the incremental run conflicting rows are completly overriden

### Expected Behavior

conflicting rows are not updated

### Steps To Reproduce

1. Have an incremental model with above config and SQL like `SELECT 'key' AS key, random() AS rand`
2. Run the model once
3. query the tabel and note down the random value
4. Run the model again
5. query table again and see that random value is different

### Relevant log output

```shell

```

### Environment

```markdown
- OS:
- Python: 3.12.13
- dbt-adapters:1.18.0
- dbt-athena-community:1.8.4
```

### Additional Context

I'm using dbt-athena with iceberg tables.

I can achieve the expected behaviour by specifying the unique_key columns in merge_update_columns.