Reactor 中的 operators publish() 和 map 之间的主要区别是什么

Question

我看过java文档，但我不知道主要的使用场景

这里是文档；

    /**
     * Share a {@link Mono} for the duration of a function that may transform it and
     * consume it as many times as necessary without causing multiple subscriptions
     * to the upstream.
     *
     * @param transform the transformation function
     * @param <R> the output value type
     *
     * @return a new {@link Mono}
     */
    public final <R> Mono<R> publish(Function<? super Mono<T>, ? extends Mono<? extends
            R>> transform) {
        return onAssembly(new MonoPublishMulticast<>(this, transform));
    }

我尝试了下面的单元测试代码片段

    @Test
    public void publish(){
        Mono<String> publish = Mono.just("test")
                .doOnNext(s -> log.info("publish --- {}",s))
                .publish(mo -> mo.map(e->e+"add String"))
                .doOnNext(s -> log.info("publish --- {}",s))
                ;

        Mono<String> map = Mono.just("test")
                .doOnNext(s -> log.info("publish --- {}", s))
                .publish(mo -> mo.map(e -> e + "add String"))
                .map(s -> s + "add String");

   }

publish

似乎与

map

没有区别，它们的用法都是将指定对象转换为同一流中的另一种类型或值