LakeFormation Data Cells Filters: TIMESTAMP data type not supported for row level filter predicate

## Description

When creating a LakeFormation Data Cells Filter with a row filter expression that references a column of type `timestamp`, the API returns:

```
InvalidInputException: TIMESTAMP data type not supported for row level filter predicate.
```

This limitation is not documented in the AWS Lake Formation documentation for Data Cells Filters.

## Reproduction

```bash
aws lakeformation create-data-cells-filter --region us-east-1 --table-data '{
"DatabaseName": "my_database",
"TableName": "my_table",
"Name": "time-restricted-filter",
"TableCatalogId": "123456789012",
"ColumnWildcard": {"ExcludedColumnNames": []},
"RowFilter": {"FilterExpression": "\"timestamp\" > '\''2026-05-01 00:00:00'\''"}
}'
```

**Error:**
```
An error occurred (InvalidInputException) when calling the CreateDataCellsFilter operation: TIMESTAMP data type not supported for row level filter predicate.
```

## Context

- Table is an Apache Iceberg table in Glue Catalog
- Column `timestamp` is of Glue type `timestamp`
- Filters on STRING columns (e.g., `servicename <> 'foo'`) work correctly
- Tested with various timestamp expression syntaxes — all fail with the same error

## Expressions tested (all fail)

- `"timestamp" > '2026-05-01 00:00:00'`
- `timestamp > date_sub(current_date(), 30)`
- `timestamp > cast(now() - interval 30 days as timestamp)`
- `timestamp >= current_timestamp - interval '30' day`
- `timestamp >= current_date() - INTERVAL 30 DAYS`

## Impact

This makes it impossible to implement time-based row-level access control using Data Cells Filters — a common use case for restricting access to older data that may contain PII per data retention policies.

## Expected behavior

Row filter expressions should support TIMESTAMP columns, or at minimum this limitation should be documented clearly in the [Data Cells Filters documentation](https://docs.aws.amazon.com/lake-formation/latest/dg/data-filtering.html).

## Environment

- AWS CLI version: 2.x
- Region: us-east-1
- Table type: Apache Iceberg (Glue Catalog)
- LakeFormation hybrid access mode enabled