/// <summary>
/// Reads levels, suitable for both repetition levels and definition levels
/// </summary>
/// <param name="reader"></param>
/// <param name="maxLevel">Maximum level value, depends on level type</param>
/// <returns></returns>
private List <int> ReadLevels(BinaryReader reader, int maxLevel)
{
int bitWidth = PEncoding.GetWidthFromMaxInt(maxLevel);
var result = new List <int>();
//todo: there might be more data on larger files, therefore line below need to be called in a loop until valueCount is satisfied
RunLengthBitPackingHybridValuesReader.ReadRleBitpackedHybrid(reader, bitWidth, 0, result);
return(result);
}
private List <int> ReadDefinitionLevels(BinaryReader reader, int valueCount)
{
const int maxDefinitionLevel = 1; //todo: for nested columns this needs to be calculated properly
int bitWidth = PEncoding.GetWidthFromMaxInt(maxDefinitionLevel);
var result = new List <int>();
//todo: there might be more data on larger files, therefore line below need to be called in a loop until valueCount is satisfied
RunLengthBitPackingHybridValuesReader.ReadRleBitpackedHybrid(reader, bitWidth, 0, result, valueCount);
int maxLevel = _schema.GetMaxDefinitionLevel(_thriftChunk);
int nullCount = valueCount - result.Count(r => r == maxLevel);
if (nullCount == 0)
{
return(null);
}
return(result);
}
private List <PageTag> WriteValues(SchemaElement schema, IList values, Thrift.PageHeader ph, CompressionMethod compression)
{
var result = new List <PageTag>();
byte[] dictionaryPageBytes = null;
int dictionaryPageCount = 0;
byte[] dataPageBytes;
List <int> repetitions = null;
List <int> definitions = null;
//flatten values and create repetitions list if the field is repeatable
if (schema.MaxRepetitionLevel > 0)
{
var rpack = new RepetitionPack();
values = rpack.Unpack(_schema, values, out repetitions);
ph.Data_page_header.Num_values = values.Count;
}
if (schema.IsNullable || schema.MaxDefinitionLevel > 0)
{
var dpack = new DefinitionPack();
values = dpack.Unpack(values, _schema, out definitions);
}
using (var ms = new MemoryStream())
{
using (var writer = new BinaryWriter(ms))
{
//write repetitions
if (repetitions != null)
{
int bitWidth = PEncoding.GetWidthFromMaxInt(_schema.MaxRepetitionLevel);
RunLengthBitPackingHybridValuesWriter.Write(writer, bitWidth, repetitions);
}
//write definitions
if (definitions != null)
{
int bitWidth = PEncoding.GetWidthFromMaxInt(_schema.MaxDefinitionLevel);
RunLengthBitPackingHybridValuesWriter.Write(writer, bitWidth, definitions);
}
//write data
if (!_writerOptions.UseDictionaryEncoding || !_dicWriter.Write(writer, schema, values, out IList dicValues))
{
_plainWriter.Write(writer, schema, values, out IList plainExtra);
}
else
{
dictionaryPageCount = dicValues.Count;
ph.Data_page_header.Encoding = Thrift.Encoding.PLAIN_DICTIONARY;
using (var dms = new MemoryStream())
using (var dwriter = new BinaryWriter(dms))
{
_plainWriter.Write(dwriter, schema, dicValues, out IList t0);
dictionaryPageBytes = dms.ToArray();
}
}
dataPageBytes = ms.ToArray();
}
}
if (dictionaryPageBytes != null)
{
Thrift.PageHeader dph = _meta.CreateDictionaryPage(dictionaryPageCount);
dictionaryPageBytes = Compress(dph, dictionaryPageBytes, compression);
int dictionaryHeaderSize = Write(dph, dictionaryPageBytes);
result.Add(new PageTag {
HeaderSize = dictionaryHeaderSize, HeaderMeta = dph
});
}
dataPageBytes = Compress(ph, dataPageBytes, compression);
int dataHeaderSize = Write(ph, dataPageBytes);
result.Add(new PageTag {
HeaderSize = dataHeaderSize, HeaderMeta = ph
});
return(result);
}
