Avro:序列化/反序列化包含Enum值的文件时发生ClassCastException

时间:2018-09-07 15:36:26

标签: java serialization deserialization avro

当我尝试反序列化以前序列化的文件时,出现以下错误:

Exception in thread "main" java.lang.ClassCastException: 
com.ssgg.bioinfo.effect.Sample$Zygosity cannot be cast to 
com.ssgg.bioinfo.effect.Sample$.Zygosity
at com.ssgg.ZygosityTest.deserializeZygosityToAvroStructure(ZygosityTest.java:45)
at com.ssgg.ZygosityTest.main(ZygosityTest.java:30)

为了重现该错误,主类如下:

public class ZygosityTest {

public static void main(String[] args) throws JsonParseException, JsonMappingException, IOException {

    String filepath = "/home/XXXX/zygosity.avro";

    /* Populate Zygosity*/
    com.ssgg.bioinfo.effect.Sample$.Zygosity zygosity = com.ssgg.bioinfo.effect.Sample$.Zygosity.HET;

    /* Create file serialized */
    createZygositySerialized(zygosity, filepath);

    /* Deserializae file */
    com.ssgg.bioinfo.effect.Sample$.Zygosity avroZygosityOutput = deserializeZygosityToAvroStructure(filepath);

}

private static com.ssgg.bioinfo.effect.Sample$.Zygosity deserializeZygosityToAvroStructure(String filepath)
        throws IOException {
    com.ssgg.bioinfo.effect.Sample$.Zygosity zygosity = null;

    File myFile = new File(filepath);
    DatumReader<com.ssgg.bioinfo.effect.Sample$.Zygosity> reader = new SpecificDatumReader<com.ssgg.bioinfo.effect.Sample$.Zygosity>(
            com.ssgg.bioinfo.effect.Sample$.Zygosity.class);
    DataFileReader<com.ssgg.bioinfo.effect.Sample$.Zygosity> dataFileReader = new DataFileReader<com.ssgg.bioinfo.effect.Sample$.Zygosity>(
            myFile, reader);

    while (dataFileReader.hasNext()) {
        zygosity = dataFileReader.next(zygosity);
    }

    dataFileReader.close();
    return zygosity;
}

private static void createZygositySerialized(com.ssgg.bioinfo.effect.Sample$.Zygosity zygosity, String filepath)
        throws IOException {
    DatumWriter<com.ssgg.bioinfo.effect.Sample$.Zygosity> datumWriter = new SpecificDatumWriter<com.ssgg.bioinfo.effect.Sample$.Zygosity>(
            com.ssgg.bioinfo.effect.Sample$.Zygosity.class);
    DataFileWriter<com.ssgg.bioinfo.effect.Sample$.Zygosity> fileWriter = new DataFileWriter<com.ssgg.bioinfo.effect.Sample$.Zygosity>(
            datumWriter);

    Schema schema = com.ssgg.bioinfo.effect.Sample$.Zygosity.getClassSchema();

    fileWriter.create(schema, new File(filepath));
    fileWriter.append(zygosity);
    fileWriter.close();
}

}

Avro为Zygosity生成的枚举如下:

/**
* Autogenerated by Avro
*
* DO NOT EDIT DIRECTLY
*/
package com.ssgg.bioinfo.effect.Sample$;
@SuppressWarnings("all")
@org.apache.avro.specific.AvroGenerated
public enum Zygosity {
  HOM_REF, HET, HOM_VAR, HEMI, UNK  ;
  public static final org.apache.avro.Schema SCHEMA$ = new     org.apache.avro.Schema.Parser().parse("{\"type\":\"enum\",\"name\":\"Zygosity\",\"namespace\":\"com.ssgg.bioinfo.effect.Sample$\",\"symbols\":[\"HOM_REF\",\"HET\",\"HOM_VAR\",\"HEMI\",\"UNK\"]}");
  public static org.apache.avro.Schema getClassSchema() { return SCHEMA$; }

}

我是Avro的新手,有人可以帮我发现问题吗? 在我的项目中,我尝试对更大的结构进行序列化和反序列化,但是我在枚举方面遇到了问题,因此我在这里隔离了一个较小的问题。 如果您需要更多信息,可以将其发布。 谢谢。

1 个答案:

答案 0 :(得分:1)

我相信这里的主要问题是$在Java类中具有特殊含义,而不太重要的是软件包名称通常是小写的。

因此,您至少应编辑名称空间以删除$