@ConfigurationProperties绑定多点点分隔属性时整后缀作为Map键问题咨询

问题原因 #

Spring Boot的@ConfigurationProperties默认会将配置项中带.的key按分隔符拆分为嵌套Map结构，因此你直接通过完整属性名从properties中取值会返回null。

解决方案 #

方案1：框架内部自动展平嵌套Map（推荐，无侵入）

无需调整使用者的配置写法，只需要修改你的CustomEclipseLinkProperties类，新增嵌套Map展平逻辑即可，后续新增任何EclipseLink属性都自动支持，不需要额外调整代码：

import org.springframework.boot.context.properties.ConfigurationProperties;
import javax.annotation.PostConstruct;
import java.util.HashMap;
import java.util.Map;
import java.util.LinkedHashMap;
import org.eclipse.persistence.config.PersistenceUnitProperties;

@ConfigurationProperties(prefix = "our.framework.eclipselink")
public class CustomEclipseLinkProperties {

    private Map<String, Object> properties;
    private Map<String, Object> flattenedProperties;

    public Map<String, Object> getProperties() {
        return properties;
    }

    public void setProperties(Map<String, Object> properties) {
        this.properties = properties;
    }

    public String getBatchSize() {
        return (String) flattenedProperties.get(PersistenceUnitProperties.BATCH_WRITING_SIZE);
    }

    @PostConstruct
    public void initFlattenedProperties() {
        flattenedProperties = new HashMap<>();
        flattenRecursive(properties, "", flattenedProperties);
    }

    private void flattenRecursive(Map<String, Object> source, String prefix, Map<String, Object> target) {
        if (source == null) return;
        for (Map.Entry<String, Object> entry : source.entrySet()) {
            String fullKey = prefix.isEmpty() ? entry.getKey() : prefix + "." + entry.getKey();
            Object value = entry.getValue();
            if (value instanceof LinkedHashMap) {
                flattenRecursive((Map<String, Object>) value, fullKey, target);
            } else {
                target.put(fullKey, value);
            }
        }
    }

    // 注入EntityManager时直接调用该方法获取所有属性即可
    public Map<String, Object> getFinalEclipseLinkProperties() {
        return flattenedProperties;
    }
}

方案2：调整配置写法

如果可以接受修改配置格式，直接在配置中把带.的属性名用方括号括起来，Spring就不会拆分key为嵌套结构：

our.framework.eclipselink.properties.[eclipselink.logging.level]=FINEST
our.framework.eclipselink.properties.[eclipselink.logging.level.cache]=FINEST
our.framework.eclipselink.properties.[eclipselink.logging.level.sql]=FINEST
our.framework.eclipselink.properties.[eclipselink.logging.parameters]=true

our.framework.eclipselink.properties.[eclipselink.jdbc.batch-writing.size]=1000
our.framework.eclipselink.properties.[eclipselink.jdbc.bind-parameters]=true
our.framework.eclipselink.properties.[eclipselink.jdbc.batch-writing]=JDBC

our.framework.eclipselink.properties.[eclipselink.jpa.uppercase-column-names]=false
our.framework.eclipselink.properties.[eclipselink.jpa.uppercase-columns]=false

调整后原getBatchSize()方法不需要修改即可正常取到值。

内容的提问来源于stack exchange，提问作者payne