// Add the size of field as a byte[] containing the 4 bytes of the integer byte size (high order byte first; char = 2 bytes)
// Read just the size -- caller must skip the field content to continue reading fields
// Return the size in bytes or chars, depending on field type
private int AddFieldSize(Document doc, FieldInfo fi, bool binary, bool compressed)
{
int size = fieldsStream.ReadVInt(), bytesize = binary || compressed?size:2 * size;
byte[] sizebytes = new byte[4];
sizebytes[0] = (byte) (SupportClass.Number.URShift(bytesize, 24));
sizebytes[1] = (byte) (SupportClass.Number.URShift(bytesize, 16));
sizebytes[2] = (byte) (SupportClass.Number.URShift(bytesize, 8));
sizebytes[3] = (byte) bytesize;
doc.Add(new Field(fi.name, sizebytes, Field.Store.YES));
return size;
}
// in merge mode we don't uncompress the data of a compressed field
private void AddFieldForMerge(Document doc, FieldInfo fi, bool binary, bool compressed, bool tokenize)
{
System.Object data;
if (binary || compressed)
{
int toRead = fieldsStream.ReadVInt();
byte[] b = new byte[toRead];
fieldsStream.ReadBytes(b, 0, b.Length);
data = b;
}
else
{
data = fieldsStream.ReadString();
}
doc.Add(new FieldForMerge(data, fi, binary, compressed, tokenize));
}
private void AddField(Document doc, FieldInfo fi, bool binary, bool compressed, bool tokenize)
{
//we have a binary stored field, and it may be compressed
if (binary)
{
int toRead = fieldsStream.ReadVInt();
byte[] b = new byte[toRead];
fieldsStream.ReadBytes(b, 0, b.Length);
if (compressed)
doc.Add(new Field(fi.name, Uncompress(b), Field.Store.COMPRESS));
else
doc.Add(new Field(fi.name, b, Field.Store.YES));
}
else
{
Field.Store store = Field.Store.YES;
Field.Index index = GetIndexType(fi, tokenize);
Field.TermVector termVector = GetTermVectorType(fi);
AbstractField f;
if (compressed)
{
store = Field.Store.COMPRESS;
int toRead = fieldsStream.ReadVInt();
byte[] b = new byte[toRead];
fieldsStream.ReadBytes(b, 0, b.Length);
f = new Field(fi.name, false, System.Text.Encoding.GetEncoding("UTF-8").GetString(Uncompress(b)), store, index, termVector);
f.SetOmitTermFreqAndPositions(fi.omitTermFreqAndPositions);
f.SetOmitNorms(fi.omitNorms);
}
else
{
f = new Field(fi.name, false, fieldsStream.ReadString(), store, index, termVector);
f.SetOmitTermFreqAndPositions(fi.omitTermFreqAndPositions);
f.SetOmitNorms(fi.omitNorms);
}
doc.Add(f);
}
}
private void AddFieldLazy(Document doc, FieldInfo fi, bool binary, bool compressed, bool tokenize)
{
if (binary)
{
int toRead = fieldsStream.ReadVInt();
long pointer = fieldsStream.GetFilePointer();
if (compressed)
{
//was: doc.add(new Fieldable(fi.name, uncompress(b), Fieldable.Store.COMPRESS));
doc.Add(new LazyField(this, fi.name, Field.Store.COMPRESS, toRead, pointer, binary));
}
else
{
//was: doc.add(new Fieldable(fi.name, b, Fieldable.Store.YES));
doc.Add(new LazyField(this, fi.name, Field.Store.YES, toRead, pointer, binary));
}
//Need to move the pointer ahead by toRead positions
fieldsStream.Seek(pointer + toRead);
}
else
{
Field.Store store = Field.Store.YES;
Field.Index index = GetIndexType(fi, tokenize);
Field.TermVector termVector = GetTermVectorType(fi);
AbstractField f;
if (compressed)
{
store = Field.Store.COMPRESS;
int toRead = fieldsStream.ReadVInt();
long pointer = fieldsStream.GetFilePointer();
f = new LazyField(this, fi.name, store, toRead, pointer, binary);
//skip over the part that we aren't loading
fieldsStream.Seek(pointer + toRead);
f.SetOmitNorms(fi.omitNorms);
f.SetOmitTermFreqAndPositions(fi.omitTermFreqAndPositions);
}
else
{
int length = fieldsStream.ReadVInt();
long pointer = fieldsStream.GetFilePointer();
//Skip ahead of where we are by the length of what is stored
if (format >= FieldsWriter.FORMAT_VERSION_UTF8_LENGTH_IN_BYTES)
fieldsStream.Seek(pointer + length);
else
fieldsStream.SkipChars(length);
f = new LazyField(this, fi.name, store, index, termVector, length, pointer, binary);
f.SetOmitNorms(fi.omitNorms);
f.SetOmitTermFreqAndPositions(fi.omitTermFreqAndPositions);
}
doc.Add(f);
}
}