Java中JSONassert、JsonUnit、Jackson对比
断言工具专为测试场景设计,支持宽松模式并精准定位差异;单元工具基于断言工具和杰克逊库封装,提供更友好的断言接口;杰克逊库作为主流处理库,可通过节点树模型自定义比对逻辑,适用于业务开发中精细控制差异的场景。
在日常开发工作中,经常需要比对两组JSON数据并返回差异结果。经过调研,整理出几款实用的JSON比对工具,可根据具体场景进行选型。
核心工具对比

1. JSONassert(测试场景首选)
JSONassert 最初为测试场景量身定制,支持忽略字段顺序、数组顺序、数值精度等宽松模式。遇到差异时直接抛出异常,异常信息会清晰指出哪个路径下的哪个值出现问题。
步骤 1:引入依赖(Maven)
org.skyscreamer jsonassert 1.5.1 test
步骤 2:实操示例(比对差异并输出)
import org.skyscreamer.jsonassert.JSONAssert;
import org.skyscreamer.jsonassert.JSONCompareMode;
public class JsonAssertDemo {
public static void main(String[] args) {
// 待比对的两个JSON字符串(结构相同,值有差异)
String expectedJson = "{\"name\":\"张三\",\"age\":20,\"address\":{\"city\":\"北京\"},\"hobbies\":[\"篮球\",\"游泳\"]}";
String actualJson = "{\"name\":\"李四\",\"age\":20,\"address\":{\"city\":\"上海\"},\"hobbies\":[\"篮球\",\"跑步\"]}";
try {
// 严格模式(字段顺序、数组顺序、值完全一致才通过)
JSONAssert.assertEquals(expectedJson, actualJson, JSONCompareMode.STRICT);
} catch (AssertionError e) {
// 捕获差异并输出
System.out.println("JSON差异:\n" + e.getMessage());
}
}
}
输出结果(精准定位差异)
JSON差异:
Expected: "张三"
got: "李四"
at path $["name"]
Expected: "北京"
got: "上海"
at path $["address"]["city"]
Expected: "游泳"
got: "跑步"
at path $["hobbies"][1]
关键配置(JSONCompareMode)
- STRICT:严格模式(字段顺序、数组顺序、值必须完全一致)
- LENIENT:宽松模式(忽略字段顺序,数组顺序仍校验)
- NON_EXTENSIBLE:不允许实际JSON包含预期外的字段
- IGNORE_ARRAY_ORDER:忽略数组元素顺序
- IGNORE_VALUES:只校验结构,忽略值的差异
2. JsonUnit(更友好的测试API)
JsonUnit 基于 JSONassert 和 Jackson 封装,API 设计更简洁,差异输出更易读。支持 assertThat 风格的断言,适合 BDD 测试风格,写测试代码更加顺手。
步骤 1:引入依赖
net.javacrumbs.json-unit json-unit-core 2.38.0 test com.fasterxml.jackson.core jackson-databind 2.15.2
步骤 2:实操示例
import static net.javacrumbs.jsonunit.assertj.JsonAssertions.assertThatJson;
public class JsonUnitDemo {
public static void main(String[] args) {
String expectedJson = "{\"name\":\"张三\",\"age\":20,\"address\":{\"city\":\"北京\"}}";
String actualJson = "{\"name\":\"李四\",\"age\":21,\"address\":{\"city\":\"北京\"}}";
// 比对并输出差异(AssertJ风格)
assertThatJson(actualJson)
.isEqualTo(expectedJson)
.onFailure(failure -> System.out.println("JSON差异:\n" + failure.getMessage()));
}
}
输出结果
JSON差异:
Expected value <"张三"> but was <"李四"> at path $['name']
Expected value <20> but was <21> at path $['age']
高级用法(忽略指定字段)
// 忽略age字段的差异
assertThatJson(actualJson)
.whenIgnoringPaths("age")
.isEqualTo(expectedJson);
3. Jackson(业务开发灵活定制)
Jackson 是 Java 生态中最主流的 JSON 处理库。通过其 JsonNode 树模型,可完全自定义比对逻辑,适合业务开发中需要精细控制差异的场景。
步骤 1:引入依赖
com.fasterxml.jackson.core jackson-databind 2.15.2
步骤 2:自定义比对工具类(输出差异路径和值)
import com.fasterxml.jackson.databind.JsonNode;
import com.fasterxml.jackson.databind.ObjectMapper;
import java.util.ArrayList;
import java.util.List;
public class JacksonJsonDiff {
private static final ObjectMapper OBJECT_MAPPER = new ObjectMapper();
private static final List DIFFS = new ArrayList<>();
// 比对两个JSON字符串,返回差异列表
public static List compareJson(String expectedJson, String actualJson) throws Exception {
DIFFS.clear();
JsonNode expectedNode = OBJECT_MAPPER.readTree(expectedJson);
JsonNode actualNode = OBJECT_MAPPER.readTree(actualJson);
compareNodes("", expectedNode, actualNode);
return DIFFS;
}
// 递归比对JsonNode节点
private static void compareNodes(String path, JsonNode expected, JsonNode actual) {
// 节点类型不同
if (expected.getNodeType() != actual.getNodeType()) {
DIFFS.add(path + ":类型不一致(预期:" + expected.getNodeType() + ",实际:" + actual.getNodeType() + ")");
return;
}
// 叶子节点(值节点)
if (expected.isValueNode()) {
if (!expected.equals(actual)) {
DIFFS.add(path + ":值不一致(预期:" + expected.asText() + ",实际:" + actual.asText() + ")");
}
return;
}
// 对象节点(递归比对子字段)
if (expected.isObject()) {
expected.fieldNames().forEachRemaining(fieldName -> {
String newPath = path.isEmpty() ? "$." + fieldName : path + "." + fieldName;
JsonNode expectedChild = expected.get(fieldName);
JsonNode actualChild = actual.get(fieldName);
if (actualChild == null) {
DIFFS.add(newPath + ":实际JSON缺失该字段");
} else {
compareNodes(newPath, expectedChild, actualChild);
}
});
// 检查实际JSON是否有预期外的字段
actual.fieldNames().forEachRemaining(fieldName -> {
if (!expected.has(fieldName)) {
String newPath = path.isEmpty() "$." + fieldName : path + "." + fieldName;
DIFFS.add(newPath + ":实际JSON包含预期外的字段");
}
});
}
// 数组节点(递归比对数组元素)
if (expected.isArray()) {
int expectedSize = expected.size();
int actualSize = actual.size();
if (expectedSize != actualSize) {
DIFFS.add(path + ":数组长度不一致(预期:" + expectedSize + ",实际:" + actualSize + ")");
}
// 比对数组元素(按索引)
int maxSize = Math.max(expectedSize, actualSize);
for (int i = 0; i < maxSize; i++) {
String newPath = path + "[" + i + "]";
JsonNode expectedChild = i < expectedSize expected.get(i) : null;
JsonNode actualChild = i < actualSize actual.get(i) : null;
if (expectedChild == null) {
DIFFS.add(newPath + ":实际数组多出元素");
} else if (actualChild == null) {
DIFFS.add(newPath + ":实际数组缺失元素");
} else {
compareNodes(newPath, expectedChild, actualChild);
}
}
}
}
// 测试
public static void main(String[] args) throws Exception {
String expectedJson = "{\"name\":\"张三\",\"age\":20,\"address\":{\"city\":\"北京\"},\"hobbies\":[\"篮球\",\"游泳\"]}";
String actualJson = "{\"name\":\"李四\",\"age\":20,\"address\":{\"city\":\"上海\"},\"hobbies\":[\"篮球\",\"跑步\"]}";
List diffs = compareJson(expectedJson, actualJson);
System.out.println("JSON差异列表:");
diffs.forEach(diff -> System.out.println("- " + diff));
}
}
输出结果
JSON差异列表:
- $.name:值不一致(预期:张三,实际:李四)
- $.address.city:值不一致(预期:北京,实际:上海)
- $.hobbies[1]:值不一致(预期:游泳,实际:跑步)
4. Diffuse(轻量级结构化差异输出)
Diffuse 是一款轻量级工具,特色在于输出 JSON 格式的结构化差异报告,将添加、修改、删除的节点逐一列出。适合需要将差异持久化或传输给其他系统的场景。
步骤 1:引入依赖
me.snov diffuse 0.2.0
步骤 2:实操示例
import me.snov.diffuse.Diffuse;
import me.snov.diffuse.JsonDiff;
public class DiffuseDemo {
public static void main(String[] args) {
String expectedJson = "{\"name\":\"张三\",\"age\":20}";
String actualJson = "{\"name\":\"李四\",\"age\":21,\"gender\":\"男\"}";
// 生成结构化差异报告
JsonDiff diff = Diffuse.diff(expectedJson, actualJson);
System.out.println("结构化差异报告:\n" + diff.toJson());
}
}
输出结果(JSON格式)
{
"changed": [
{
"path": "/name",
"from": "张三",
"to": "李四"
},
{
"path": "/age",
"from": 20,
"to": 21
}
],
"added": [
{
"path": "/gender",
"value": "男"
}
],
"removed": []
}
综上所述,选型并不复杂,可根据实际需求决定:
- 测试场景:优先选 JsonUnit(API 友好)或 JSONassert(轻量高效);
- 业务开发:优先选 Jackson(灵活自定义比对规则);
- 需要结构化差异输出:选 Diffuse;
- 特殊需求(如忽略数值精度、时间格式):基于 Jackson 扩展自定义比对逻辑。
