[VL] Scan split proto contains duplicate field

### Description

```json
{
"uriFile": "oss://xxx/tpcds/partitioned/tpcds_sf100/web_sales/ws_sold_date_sk=2451145/part-00086-5f7cfd3c-f467-4d40-82e3-d2033487db84.c000.snappy.parquet",
"partitionIndex": "5",
"length": "5483534",
"parquet": {},
"partitionColumns": [
{
"key": "ws_sold_date_sk",
"value": "2451145"
}
],
"schema": {},
"metadataColumns": [
{
"key": "input_file_name",
"value": "oss://xxx/tpcds/partitioned/tpcds_sf100/web_sales/ws_sold_date_sk=2451145/part-00086-5f7cfd3c-f467-4d40-82e3-d2033487db84.c000.snappy.parquet"
},
{
"key": "input_file_block_length",
"value": "5483534"
},
{
"key": "input_file_block_start",
"value": "0"
}
],
"properties": {
"fileSize": "5483534",
"modificationTime": "1719801698242"
}
}
```

some fields are duplicate, should be merged.

@gaoyangxiaozhu @acvictor

Start with the scan split proto output shown in the issue and trace where its fields are assembled or serialized. No files or tests are named; done means determining the intended merged representation and confirming that duplicate fields no longer appear in the scan split.

Written by the indexing model from the issue text.