Fix merge into compatability

[osipovartem]

Based on Embucket/iceberg-rust#52 and
Embucket/iceberg-rust#53

• for overwrite operation we need to save filtered data files for compatability with other iceberg engines since they take into account all deteled data files and stats within current snapshot
  Spark andSnowflake create additional manifest to track all deleted files for compatability

  manifest_path: s3://...-m1.avro
  manifest_length: 7944
  partition_spec_id: 0
  content: 0
  sequence_number: 2
  min_sequence_number: 2
  added_snapshot_id: 8234522747398339109
  added_files_count: 1
  existing_files_count: 0
  deleted_files_count: 0
  added_rows_count: 6001215
  existing_rows_count: 0
  deleted_rows_count: 0
  partitions: []
  key_metadata: None

Record 2:
  manifest_path: s3://...-m0.avro
  manifest_length: 7942
  partition_spec_id: 0
  content: 0
  sequence_number: 2
  min_sequence_number: 2
  added_snapshot_id: 8234522747398339109
  added_files_count: 0
  existing_files_count: 0
  deleted_files_count: 1
  added_rows_count: 0
  existing_rows_count: 0
  deleted_rows_count: 6001215
  partitions: []
  key_metadata: None

• this mechanism can be also used in future to restore removed data