java - 使用 Jackson 解析数组

Question

我想使用 Jackson 解析以下 Json，但我完全不确定应该如何构建实体。

   [
      CHAN_ID, 
      [
        [
          SYMBOL, 
          STATUS, 
          AMOUNT, 
          BASE_PRICE, 
          MARGIN_FUNDING, 
          MARGIN_FUNDING_TYPE,
          PL,
          PL_PERC,
          PRICE_LIQ,
          LEVERAGE,
           ...
        ], 
        ...
      ]
    ]

这是我到目前为止所拥有的:

@Getter
@Setter
@EqualsAndHashCode
@ToString
@NoArgsConstructor
public class Position {

        @JsonProperty("SYMBOL")
        private String symbol;
        @JsonProperty("STATUS")
        private String status;
        @JsonProperty("AMOUNT")
        private Decimal amount;
        @JsonProperty("BASE_PRICE")
        private Decimal basePrice;
        @JsonProperty("MARGIN_FUNDING")
        private Integer marginFunding;
        @JsonProperty("MARGIN_FUNDING_TYPE")
        private Decimal marginFundingType;
        @JsonProperty("PL")
        private Decimal profitLoss;
        @JsonProperty("PL_PERC")
        private Decimal profitLossPercentage;
        @JsonProperty("PRICE_LIQ")
        private Decimal liquidationPrice;
        @JsonProperty("LEVERAGE")
        private Decimal leverage;
    }

我试图解析的这个东西，似乎有一个 Position 数组，但在它有 CHAN_ID 之前，我应该为此构造某种包装类吗？

@Getter
@Setter
@EqualsAndHashCode
@ToString
@NoArgsConstructor
public class Positions {

    @JsonProperty("CHAN_ID")
    private String channelId;

    @JsonProperty("positions")
    private List<Position> positions;

}

您认为这是否正确？也不介意类顶部的那些注释，它只是 Lombok。目前，当我尝试解析时，此实现会出现以下错误:

com.fasterxml.jackson.databind.exc.MismatchedInputException: Cannot deserialize instance of model.Positions out of START_ARRAY token at [Source: (String)"['ps', [ [ 'aa', 'bb', 123.45, 123.45, 567, 123.45, 123.45, 123.45, 123.45, 123.45]] ]"; line: 1, column: 1]

为了测试这一点，我只使用 ObjectMapper:

public static void main(String[] args) throws IOException {
        ObjectMapper objectMapper = new ObjectMapper();
        try {
            Positions positions = objectMapper.readValue("['ps', [ [ 'aa', 'bb', 123.45, 123.45, 567, 123.45, 123.45, 123.45, 123.45, 123.45]] ]", Positions.class);
        } catch (IOException e) {
            e.printStackTrace();
        }

    }

Answer 1

由于输入 JSON 的结构为嵌套数组，因此与 @JsonProperty 绑定(bind)没有多大意义，因为输入 JSON 中不存在这些属性键。仅当您想随后将反序列化数据转换为具有属性值对的 JSON 时，它们才有意义。

要正确读取此 JSON 结构，您必须实现自定义反序列化器，因为在这种情况下，属性无法自动与其值进行映射。反序列化器可以是 Jackson 的 JsonDeserializer 或 StdDeserializer 的实现。例如，以下实现有效:

import com.fasterxml.jackson.core.JsonParser;
import com.fasterxml.jackson.databind.DeserializationContext;
import com.fasterxml.jackson.databind.deser.std.StdDeserializer;

import java.io.IOException;
import java.util.List;
import java.util.stream.Collectors;

public class PositionsDeserializer extends StdDeserializer<Positions> {

  public PositionsDeserializer() {
    super(Positions.class);
  }

  @Override
  public Positions deserialize(JsonParser p, DeserializationContext ctxt) throws IOException {
    Object[] positionsProps = p.readValueAs(Object[].class);
    if (positionsProps != null && positionsProps.length > 0) {
      String chanId = (String) positionsProps[0];
      List<Position> positionsList =
          ((List<List<?>>) positionsProps[1]).stream()
              .map(this::deserializePosition)
              .collect(Collectors.toList());
      Positions positions = new Positions();
      positions.setChannelId(chanId);
      positions.setPositions(positionsList);
      return positions;
    }
    // decide whether you want to return null, throw an exception or other outcome: depends on the constraints of your data
    return null;
  }

  private Position deserializePosition(List<?> props) {
    if (props != null && !props.isEmpty()) {
      final Position position = new Position();
      position.setSymbol((String) props.get(0));
      position.setStatus((String) props.get(1));
      position.setAmount((Double) props.get(2));
      position.setBasePrice((Double) props.get(3));
      position.setMarginFunding((Integer) props.get(4));
      position.setMarginFundingType((Double) props.get(5));
      position.setProfitLoss((Double) props.get(6));
      position.setProfitLossPercentage((Double) props.get(7));
      position.setLiquidationPrice((Double) props.get(8));
      position.setLeverage((Double) props.get(9));
      return position;
    }
    // decide whether you want to return null, throw an exception or other outcome: depends on the constraints of your data
    return null;
  }

}

要使用此反序列化器，应在 ObjectMapper 中注册它:

private ObjectMapper initObjectMapper() {
  ObjectMapper objectMapper = new ObjectMapper();
  SimpleModule module = new SimpleModule();
  module.addDeserializer(Positions.class, new PositionsDeserializer());
  objectMapper.registerModule(module);
  return objectMapper;
}

@Test
public void deserializePositions() throws IOException {
  String json = "[ \"ps\", [ [ \"aa\", \"bb\", 123.45, 123.45, 567, 123.45, 123.45, 123.45, 123.45, 123.45 ] ] ]";

  Positions positions = initObjectMapper().readValue(json, Positions.class);

  System.out.println(positions);
}

建议

如果您可以选择/更改输入 JSON 的结构，我建议减少嵌套并引入更传统的 KV 映射。例如，要表示位置:

{
  "CHAN_ID": "string",
  "positions": [
    // Use the arrays here: [ "aa", "bb", 123.45, 123.45, 567, 123.45, 123.45, 123.45, 123.45, 123.45 ], 
    // OR transform the array into KV mapping: { "SYMBOL": "string", ... }, ...
  ]
}

这种方法将允许:

• 减少由与值的数组索引相关的魔数(Magic Number)引起的错误风险；

• 简化反序列化:自动读取Positions，并仅在保留数组结构的情况下为Position提供自定义反序列化器。

另外，我不知道原始代码中的 Decimal 类型是什么意思，所以我将其替换为 Double。