Side-effects only when using Pipeable Operators?

Question

I'm rewriting some of my code to use Pipeable Operators in place of "patch" operators and I'm a little confused about when to use pipe() and how do I subscribe to a piped Observable to "activate" it if I only want side-effects?

For example, I have this code:

this.messages$ = this.messageService.getMessages(messageType)
    .do((messages: Message[]) => {
        console.log('The first message is ' + deviceMessages[0].text);
    });

I get confused here because I figure I want to do this.messageService.getMessages(messageType).pipe(...) so I get an Observable returned and assigned to this.messages$ but then that won't execute because I haven't subscribe()'d. But if I do .pipe(...).subscribe() then it'll return a Subscription object, which isn't what I want.

What should I be doing here?

Answer 1

Well, do it simply like this:

this.messages$ = this.messageService.getMessages(messageType).pipe(
  tap((messages: Message[]) => {
    console.log('The first message is ' + deviceMessages[0].text);
  })
);
this.messages$.subscribe();

the tap operator is the equivalent of .do() in rxjs 5 and above.