Java introspection: object to map

Question

I have a Java object obj that has attributes obj.attr1, obj.attr2 etc. The attributes are possibly accessed through an extra level of indirection: obj.getAttr1(), obj.getAttr2(), if not public.

The challenge: I want a function that takes an object, and returns a Map<String, Object>, where the keys are strings "attr1", "attr2" etc. and values are the corresponding objects obj.attr1, obj.attr2. I imagine the function would be invoked with something like

• toMap(obj),
• or toMap(obj, "attr1", "attr3") (where attr1 and attr3 are a subset of obj's attributes),
• or perhaps toMap(obj, "getAttr1", "getAttr3") if necessary.

I don't know much about Java's introspection: how do you do that in Java?

Right now, I have a specialized toMap() implementation for each object type that I care about, and it's too much boilerplate.

---

NOTE: for those who know Python, I want something like obj.__dict__. Or dict((attr, obj.__getattribute__(attr)) for attr in attr_list) for the subset variant.

Answer 1

Another way to user JacksonObjectMapper is the convertValue ex:

 ObjectMapper m = new ObjectMapper();
 Map<String,Object> mappedObject = m.convertValue(myObject, new TypeReference<Map<String, String>>() {});

Answer 2

Use Apache Commons BeanUtils: http://commons.apache.org/beanutils/.

An implementation of Map for JavaBeans which uses introspection to get and put properties in the bean:

Map<Object, Object> introspected = new org.apache.commons.beanutils.BeanMap(object); 

Note: despite the fact the API returns Map<Object, Object> (since 1.9.0), the actual class for keys in the returned map is java.lang.String

Answer 3

You can use JavaBeans introspection for this. Read up on the java.beans.Introspector class:

public static Map<String, Object> introspect(Object obj) throws Exception {
    Map<String, Object> result = new HashMap<String, Object>();
    BeanInfo info = Introspector.getBeanInfo(obj.getClass());
    for (PropertyDescriptor pd : info.getPropertyDescriptors()) {
        Method reader = pd.getReadMethod();
        if (reader != null)
            result.put(pd.getName(), reader.invoke(obj));
    }
    return result;
}

Big caveat: My code deals with getter methods only; it will not find naked fields. For fields, see highlycaffeinated's answer. :-) (You will probably want to combine the two approaches.)

Answer 4

Here's a rough approximation, hopefully enough to get you pointed in the right direction:

public Map<String, Object> getMap(Object o) {
    Map<String, Object> result = new HashMap<String, Object>();
    Field[] declaredFields = o.getClass().getDeclaredFields();
    for (Field field : declaredFields) {
        result.put(field.getName(), field.get(o));
    }
    return result;
}

Answer 5

Here is a really easy way to do this.

Use Jackson JSON lib to convert the object to JSON.

Then read the JSON and convert it to a Map.

The map will contain everything you want.

Here is the 4 liner

ObjectMapper om = new ObjectMapper();
StringWriter sw = new StringWriter();
om.writeValue(object, sw);
Map<String, Object> map = om.readValue(sw.toString(), Map.class);

And additional win of course is that this is recursive and will create maps of maps if it needs to

Answer 6

None of these work for nested properties, object mapper does a fair job except that you have to set all values on all fields you want to see in map and even then you cannot avoid/ignore objects own @Json annotations easily in ObjectMapper basically skip some of the properties. So unfortunately, you have to do something like the following, it is only a draft to just give an idea.

/*
     * returns fields that have getter/setters including nested fields as
     * field0, objA.field1, objA.objB.field2, ... 
     * to take care of recursive duplicates, 
     * simply use a set<Class> to track which classes
     * have already been traversed
     */
    public static void getBeanUtilsNestedFields(String prefix, 
            Class clazz,  List<String> nestedFieldNames) throws Exception {
        PropertyDescriptor[] descriptors = BeanUtils.getPropertyDescriptors(clazz);
        for(PropertyDescriptor descr : descriptors){
            // if you want values, use: descr.getValue(attributeName)
            if(descr.getPropertyType().getName().equals("java.lang.Class")){
                continue;
            }
            // a primitive, a CharSequence(String), Number, Date, URI, URL, Locale, Class, or corresponding array
            // or add more like UUID or other types
            if(!BeanUtils.isSimpleProperty(descr.getPropertyType())){
                Field collectionfield = clazz.getDeclaredField(descr.getName());
                if(collectionfield.getGenericType() instanceof ParameterizedType){
                    ParameterizedType integerListType = (ParameterizedType) collectionfield.getGenericType();
                    Class<?> actualClazz = (Class<?>) integerListType.getActualTypeArguments()[0];
                    getBeanUtilsNestedFields(descr.getName(), actualClazz, nestedFieldNames);
                }
                else{   // or a complex custom type to get nested fields
                    getBeanUtilsNestedFields(descr.getName(), descr.getPropertyType(), nestedFieldNames);
                }
            }
            else{
                nestedFieldNames.add(prefix.concat(".").concat(descr.getDisplayName()));
            }
        }
    }

Answer 7

maven dependencies

    <dependency>
        <groupId>com.fasterxml.jackson.datatype</groupId>
        <artifactId>jackson-datatype-jsr310</artifactId>
    </dependency>

....

ObjectMapper m = new ObjectMapper();
Map<String,Object> mappedObject = m.convertValue(myObject,Map.class);

for JSR310 New Date/Time API,there are some issue need to be improved eg:

import com.alibaba.fastjson.JSON;
import com.fasterxml.jackson.databind.ObjectMapper;
import lombok.Data;
import lombok.NoArgsConstructor;
import org.junit.Test;

import java.sql.Timestamp;
import java.time.LocalDateTime;
import java.time.ZoneId;
import java.util.Map;

@Data
@NoArgsConstructor
public class QueryConditionBuilder
{
    LocalDateTime startTime;
    LocalDateTime endTime;
    Long nodeId;
    Long fsId;
    Long memId;
    Long ifCardId;

    private QueryConditionBuilder(QueryConditionBuilder.Builder builder) {
        setStartTime(builder.startTime);
        setEndTime(builder.endTime);
        setNodeId(builder.nodeId);
        setFsId(builder.fsId);
        setMemId(builder.memId);
        setIfCardId(builder.ifCardId);
    }

    public static QueryConditionBuilder.Builder newBuilder() {
        return new QueryConditionBuilder.Builder();
    }

    public static QueryConditionBuilder newEmptyBuilder() {
        return new QueryConditionBuilder.Builder().build();
    }


    public Map<String,Object> toFilter()
    {
        Map<String,Object> filter = new ObjectMapper().convertValue(this,Map.class);
        System.out.printf("查询条件:%s\n", JSON.toJSONString(filter));
        return filter;
    }

    public static final class Builder {
        private LocalDateTime startTime;
        private LocalDateTime endTime;
        private Long nodeId = null;
        private Long fsId = null;
        private Long memId =null;
        private Long ifCardId = null;

        private Builder() {
        }

        public QueryConditionBuilder.Builder withStartTime(LocalDateTime val) {
            startTime = val;
            return this;
        }

        public QueryConditionBuilder.Builder withEndTime(LocalDateTime val) {
            endTime = val;
            return this;
        }

        public QueryConditionBuilder.Builder withNodeId(Long val) {
            nodeId = val;
            return this;
        }

        public QueryConditionBuilder.Builder withFsId(Long val) {
            fsId = val;
            return this;
        }

        public QueryConditionBuilder.Builder withMemId(Long val) {
            memId = val;
            return this;
        }

        public QueryConditionBuilder.Builder withIfCardId(Long val) {
            ifCardId = val;
            return this;
        }

        public QueryConditionBuilder build() {
            return new QueryConditionBuilder(this);
        }
    }

    @Test
    public void test()
    {     
        LocalDateTime now = LocalDateTime.now(ZoneId.of("+8"));
        LocalDateTime yesterday = now.plusHours(-24);

        Map<String, Object> condition = QueryConditionBuilder.newBuilder()
                .withStartTime(yesterday)
                .withEndTime(now)
                .build().toFilter();

        System.out.println(condition);
    }
}

expects(pseudo-code):

查询条件:{"startTime":{"2019-07-15T20:43:15"},"endTime":{"2019-07-16T20:43:15"}
{startTime={2019-07-15T20:43:15}, endTime={"2019-07-16T20:43:15"}, nodeId=null, fsId=null, memId=null, ifCardId=null}

instead,i got these:

查询条件:{"startTime":{"dayOfMonth":15,"dayOfWeek":"MONDAY","dayOfYear":196,"hour":20,"minute":38,"month":"JULY","monthValue":7,"nano":263000000,"year":2019,"second":12,"chronology":{"id":"ISO","calendarType":"iso8601"}},"endTime":{"dayOfMonth":16,"dayOfWeek":"TUESDAY","dayOfYear":197,"hour":20,"minute":38,"month":"JULY","monthValue":7,"nano":263000000,"year":2019,"second":12,"chronology":{"id":"ISO","calendarType":"iso8601"}}}
{startTime={dayOfMonth=15, dayOfWeek=MONDAY, dayOfYear=196, hour=20, minute=38, month=JULY, monthValue=7, nano=263000000, year=2019, second=12, chronology={id=ISO, calendarType=iso8601}}, endTime={dayOfMonth=16, dayOfWeek=TUESDAY, dayOfYear=197, hour=20, minute=38, month=JULY, monthValue=7, nano=263000000, year=2019, second=12, chronology={id=ISO, calendarType=iso8601}}, nodeId=null, fsId=null, memId=null, ifCardId=null}

after a few research,an effective trick was found,

ObjectMapper mapper = new ObjectMapper();
JavaTimeModule module = new JavaTimeModule();
//https://github.com/networknt/light-4j/issues/82
mapper.registerModule(module);
mapper.configure(SerializationFeature.WRITE_DATES_AS_TIMESTAMPS, false);
//incase of empty/null String
mapper.setSerializationInclusion(JsonInclude.Include.NON_NULL);
mapper.setSerializationInclusion(JsonInclude.Include.NON_EMPTY);
Map<String,Object> filter = mapper.convertValue(this,Map.class);
System.out.printf("查询条件:%s\n", JSON.toJSONString(filter));
return filter;

output:

查询条件:{"startTime":"2019-07-15T21:29:13.711","endTime":"2019-07-16T21:29:13.711"}
{startTime=2019-07-15T21:29:13.711, endTime=2019-07-16T21:29:13.711, nodeId=null, fsId=null, memId=null, ifCardId=null}

I used the above code for dynamical query in MyBatis
eg.

 /***
     * 查询文件系统使用率
     * @param condition
     * @return
     */
    LinkedList<SnmpFileSystemUsage> queryFileSystemUsage(Map<String,Object> condition);

    List<SnmpFileSystemUsage> fooBar()
    { 
       return snmpBaseMapper.queryFileSystemUsage(QueryConditionBuilder
                .newBuilder()
                .withNodeId(nodeId)
                .build()
                .toFilter());
    }