/// <summary>
/// Reads entire file as a table
/// </summary>
public static Table ReadTableFromFile(string filePath, ParquetOptions parquetOptions = null)
{
using (ParquetReader reader = OpenFromFile(filePath, parquetOptions))
{
return(reader.ReadAsTable());
}
}
/// <summary>
/// Reads <see cref="DataSet"/> from an open stream
/// </summary>
/// <param name="source">Input stream</param>
/// <param name="formatOptions">Parquet options, optional.</param>
/// <param name="readerOptions">Reader options, optional</param>
/// <returns><see cref="DataSet"/></returns>
public static DataSet Read(Stream source, ParquetOptions formatOptions = null, ReaderOptions readerOptions = null)
{
using (var reader = new ParquetReader(source, formatOptions, readerOptions))
{
return(reader.Read());
}
}
/// <summary>
/// Reads entire stream as a table
/// </summary>
public static Table ReadTableFromStream(Stream stream, ParquetOptions parquetOptions = null)
{
using (var reader = new ParquetReader(stream, parquetOptions))
{
return(reader.ReadAsTable());
}
}
/// <summary>
///
/// </summary>
/// <typeparam name="T"></typeparam>
/// <param name="input"></param>
/// <returns></returns>
public static IEnumerable <T[]> DeserializeGroups <T>(Stream input) where T : new()
{
var bridge = new ClrBridge(typeof(T));
using (var reader = new ParquetReader(input))
{
Schema fileSchema = reader.Schema;
DataField[] dataFields = fileSchema.GetDataFields();
for (int i = 0; i < reader.RowGroupCount; i++)
{
using (ParquetRowGroupReader groupReader = reader.OpenRowGroupReader(i))
{
DataColumn[] groupColumns = dataFields
.Select(df => groupReader.ReadColumn(df))
.ToArray();
T[] rb = new T[groupReader.RowCount];
for (int ie = 0; ie < rb.Length; ie++)
{
rb[ie] = new T();
}
for (int ic = 0; ic < groupColumns.Length; ic++)
{
bridge.AssignColumn(groupColumns[ic], rb);
}
yield return(rb);
}
}
}
}
/// <summary>
///
/// </summary>
/// <typeparam name="T"></typeparam>
/// <param name="input"></param>
/// <param name="rowGroupIndex"></param>
/// <returns></returns>
public static T[] Deserialize <T>(Stream input, int rowGroupIndex = -1) where T : new()
{
var result = new List <T>();
using (var reader = new ParquetReader(input))
{
Schema fileSchema = reader.Schema;
DataField[] dataFields = fileSchema.GetDataFields();
if (rowGroupIndex == -1) //Means read all row groups.
{
for (int i = 0; i < reader.RowGroupCount; i++)
{
T[] currentRowGroupRecords = ReadAndDeserializeByRowGroup <T>(i, reader, dataFields);
result.AddRange(currentRowGroupRecords);
}
}
else //read specific rowgroup.
{
T[] currentRowGroupRecords = ReadAndDeserializeByRowGroup <T>(rowGroupIndex, reader, dataFields);
result.AddRange(currentRowGroupRecords);
}
}
return(result.ToArray());
}
/// <summary>
/// Reads the first row group as a table
/// </summary>
/// <param name="reader">Open reader</param>
/// <returns></returns>
public static Table ReadAsTable(this ParquetReader reader)
{
Table result = null;
for (int i = 0; i < reader.RowGroupCount; i++)
{
using (ParquetRowGroupReader rowGroupReader = reader.OpenRowGroupReader(i))
{
DataColumn[] allData = reader.Schema.GetDataFields().Select(df => rowGroupReader.ReadColumn(df)).ToArray();
var t = new Table(reader.Schema, allData, rowGroupReader.RowCount);
if (result == null)
{
result = t;
}
else
{
foreach (Row row in t)
{
result.Add(row);
}
}
}
}
return(result);
}
/// <summary>
/// Reads the file
/// </summary>
/// <param name="fullPath">The full path.</param>
/// <param name="formatOptions">Optional reader options.</param>
/// <param name="readerOptions">The reader options.</param>
/// <returns><see cref="DataSet"/></returns>
public static DataSet ReadFile(string fullPath, ParquetOptions formatOptions = null, ReaderOptions readerOptions = null)
{
using (Stream fs = System.IO.File.OpenRead(fullPath))
{
using (var reader = new ParquetReader(fs, formatOptions, readerOptions))
{
return(reader.Read());
}
}
}
/// <summary>
/// Reads the first row group from a file
/// </summary>
/// <param name="stream"></param>
/// <param name="schema"></param>
/// <param name="columns"></param>
public static void ReadSingleRowGroupParquetFile(this Stream stream, out Schema schema, out DataColumn[] columns)
{
using (var reader = new ParquetReader(stream))
{
schema = reader.Schema;
using (ParquetRowGroupReader rgr = reader.OpenRowGroupReader(0))
{
DataField[] dataFields = schema.GetDataFields();
columns = new DataColumn[dataFields.Length];
for (int i = 0; i < dataFields.Length; i++)
{
columns[i] = rgr.ReadColumn(dataFields[i]);
}
}
}
}
private static T[] ReadAndDeserializeByRowGroup <T>(int rowGroupIndex, ParquetReader reader, DataField[] dataFields) where T : new()
{
var bridge = new ClrBridge(typeof(T));
using (ParquetRowGroupReader groupReader = reader.OpenRowGroupReader(rowGroupIndex))
{
DataColumn[] groupColumns = dataFields
.Select(df => groupReader.ReadColumn(df))
.ToArray();
T[] rb = new T[groupReader.RowCount];
for (int ie = 0; ie < rb.Length; ie++)
{
rb[ie] = new T();
}
for (int ic = 0; ic < groupColumns.Length; ic++)
{
bridge.AssignColumn(groupColumns[ic], rb);
}
return(rb);
}
}
