Skip empty files while migrating tables to iceberg

### Feature Request / Improvement

ORC:

```
java.lang.IllegalArgumentException: ORC schema does not contain Iceberg IDs
at org.apache.iceberg.orc.ORCSchemaUtil.convert(ORCSchemaUtil.java:293)
at org.apache.iceberg.orc.OrcMetrics.buildOrcMetrics(OrcMetrics.java:143)
at org.apache.iceberg.orc.OrcMetrics.fromInputFile(OrcMetrics.java:93)
at org.apache.iceberg.orc.OrcMetrics.fromInputFile(OrcMetrics.java:87)
at org.apache.iceberg.data.TableMigrationUtil.getOrcMetrics(TableMigrationUtil.java:296)
at org.apache.iceberg.data.TableMigrationUtil.lambda$listPartition$4(TableMigrationUtil.java:244)
```

Parquet:

```
java.lang.RuntimeException: mig_parquet_sample/date_key=2026-09-02/1_$folder$ is not a Parquet file (length is too low: 0)
at org.apache.iceberg.shaded.org.apache.parquet.hadoop.ParquetFileReader.readFooter(ParquetFileReader.java:599)
at org.apache.iceberg.shaded.org.apache.parquet.hadoop.ParquetFileReader.readFooter(ParquetFileReader.java:578)
at org.apache.iceberg.shaded.org.apache.parquet.hadoop.ParquetFileReader.(ParquetFileReader.java:971)
at org.apache.iceberg.shaded.org.apache.parquet.hadoop.ParquetFileReader.(ParquetFileReader.java:961)
at org.apache.iceberg.shaded.org.apache.parquet.hadoop.ParquetFileReader.open(ParquetFileReader.java:718)
at org.apache.iceberg.parquet.ParquetUtil.fileMetrics(ParquetUtil.java:64)
at org.apache.iceberg.data.TableMigrationUtil.getParquetMetrics(TableMigrationUtil.java:287)
at org.apache.iceberg.data.TableMigrationUtil.lambda$listPartition$3(TableMigrationUtil.java:236)

```

Sometimes empty files are created in hive, we need to manually remove all empty files before migrating to iceberg.

We can fix this by skipping empty files

### Query engine

Spark

### Willingness to contribute

- [x] I can contribute this improvement/feature independently
- [x] I would be willing to contribute this improvement/feature with guidance from the Iceberg community
- [ ] I cannot contribute this improvement/feature at this time

Read TableMigrationUtil, especially listPartition and its getOrcMetrics and getParquetMetrics paths shown in the stack traces. Trace how migration handles discovered files, then verify that zero-length files are skipped before ORC or Parquet metrics are read; done means migrations no longer fail on empty files.

Written by the indexing model from the issue text.