在使用protobuf-net对c#中的大文件(20GB)进行反序列化时,我得到"Sub-message not read correctly" Exception
。读取2 GB的20 GB后发生异常。
相同的数据结构适用于较小的实例。
20 MB数据的序列化工作正常。
以下是序列化的一些示例代码:
if (File.Exists(filename))
File.Delete(filename);
using (FileStream stream = new FileStream(filename, FileMode.Create))
{
Serializer.Serialize<HubLabelingData>(stream, data);
stream.Close();
}
以下是反序列化的一些示例代码:
using (FileStream stream = new FileStream(filename, FileMode.Open))
{
data = Serializer.Deserialize<HubLabelingData>(stream);
stream.Close();
}
这是数据结构(适用于小实例):
[ProtoContract]
public class HubLabelingData
{
[ProtoMember(1)]
public HL[] hlf;
[ProtoMember(2)]
public HL[] hlb;
[ProtoMember(3)]
public NG g;
[ProtoMember(4)]
public List<PL> plf;
[ProtoMember(5)]
public List<PL> plb;
[ProtoMember(6)]
public PHL[] pihlf;
[ProtoMember(7)]
public PHL[] pihlb;
}
[ProtoContract]
public class HL
{
[ProtoMember(1)]
public int[] l;
[ProtoMember(2)]
public double[] d;
}
[ProtoContract]
public class PL
{
[ProtoMember(1)]
public Dictionary<int, List<GP>> p;
}
[ProtoContract]
public class PHL
{
[ProtoMember(1)]
public short[] l;
}
[ProtoContract]
public class NG
{
[ProtoMember(1)]
public NA[] e;
[ProtoMember(2)]
public NA[] tne { get; set; }
[ProtoMember(3)]
public float[] a;
[ProtoMember(4)]
public float[] o;
[ProtoMember(5)]
public int num = 0;
}
[ProtoContract]
public class NA
{
[ProtoMember(1)]
public int one { get; set; }
[ProtoMember(2)]
public int two { get; set; }
[ProtoMember(3)]
public double tree { get; set; }
[ProtoMember(4)]
public int four { get; set; }
}
[ProtoContract]
public class NN
{
[ProtoMember(1)]
public List<NA> nas;
}
[ProtoContract]
public class GP
{
[ProtoMember(1)]
public float one { get; set; }
[ProtoMember(2)]
public float two { get; set; }
}
答案 0 :(得分:1)
对于遇到同样问题的每个人:
Protbuf.net不支持反序列化大于2.048 GB的文件(分别为对象)。
我将数据拆分成多条消息。更多的工作,但它的工作完美。
答案 1 :(得分:0)
protobuf-net支持从v2.2.0开始对大于2GB的流进行反序列化,但有警告。
从v2.0.0.668迁移到v2.3.13对我来说已经解决了此问题,以前我遇到了"Sub-message not read correctly"
异常。