protected override void OnBeginFile()
{
// Switching to a new input file, so reset our read variables.
var options = new Parquet.ParquetOptions();
_pqReader = new Parquet.ParquetReader(_reader.BaseStream);
_pqDataFields = _pqReader.Schema.GetDataFields();
_rowGroupIndex = 0;
_rowGroupReadCount = 0;
_rowGroupRowCount = 0;
}
protected override void OnBeginReading()
{
// Use Parquet.NET to do the heavy reading.
var options = new Parquet.ParquetOptions();
_pqReader = new Parquet.ParquetReader(_reader.BaseStream);
_pqDataFields = _pqReader.Schema.GetDataFields();
// Extract the column names from the parquet file in a separate array for fast lookups.
for (int i = 0; i < _pqDataFields.Length; i++)
{
var columnSchema = new ColumnSchema()
{
ColumnName = _pqDataFields[i].Name,
DataTypeName = _pqDataFields[i].DataType.ToString(),
ColumnSize = -1,
NumericPrecision = -1,
NumericScale = -1,
AllowNull = _pqDataFields[i].HasNulls
};
_columnSchema.Add(columnSchema);
}
}
public MetadataViewer(Parquet.ParquetReader parquetReader) : this()
{
this.parquetReader = parquetReader;
}
