fix: check mayContainNulls in ParquetDictionaryRowGroupFilter.notStartsWith

[waterWang]

Fixes #17655

Problem

ParquetDictionaryRowGroupFilter.notStartsWith can skip row groups that contain null values, silently dropping rows from scan results.

The filter decides whether a row group can be skipped by inspecting only the column's dictionary. A Parquet dictionary contains non-null values only, so when every dictionary entry starts with the prefix the filter returns ROWS_CANNOT_MATCH — even when the column also contains nulls.

In Iceberg a null value matches notStartsWith: Evaluator implements it as !startsWith(...), and startsWith evaluates to false for null. A row group holding nulls therefore does contain matching rows and must not be skipped.

Fix

Add a mayContainNulls check to notStartsWith, consistent with how notEq, notIn, and notNaN already handle null-containing row groups.

Changes

1. ParquetDictionaryRowGroupFilter.java: Added if (mayContainNulls.get(id)) return ROWS_MIGHT_MATCH; in notStartsWith, matching the pattern used by notNaN.
2. TestDictionaryRowGroupFilter.java: Updated the assertion for notStartsWith("some_nulls", "some") from isFalse() to isTrue() — this row group contains nulls, so it must be read.

[uros-b]

I think this is already in progress #17656.