feat: support reading partitioned Delta table.

[youngsofun]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

In a delta table, partition columns are not stored in parquet file.
so it needs a few efforts to make pushdown work:

• context:
  • Table store partition column names in meta.engine_options.
  • Each partition carries all partition column values in the same order.
  • With this order, we can get needed info with a PartitionIndex.
• pushdown:
  • projections (mask): partition columns are excluded when read parquet file and inserted at last. (use FieldIndex->PartitionIndex)
  • filter pass to parquet reader: all partition columns are appended to the filter input columns. （ID=PartitionIndex+ num_of_non_partition_input_columns）
  • pruner: ColumnRef of partition columns in filter expr are replace with const scalars (by name).

Type of partition columns can only be simple primitive types.

Tests

• Unit Test
• Logic Test
• Benchmark Test
• No Test - Explain why

Type of change

• Bug Fix (non-breaking change which fixes an issue)
• New Feature (non-breaking change which adds functionality)
• Breaking Change (fix or feature that could cause existing functionality not to work as expected)
• Documentation Update
• Refactoring
• Performance Improvement
• Other (please describe):

---

This change is