Hacker News new | past | comments | ask | show | jobs | submit login

I've seen and understand Redux TO DO examples.

However I develop enterprise CRM app. In db there are 200k client records, 500k sales calls records. It is implemented as a standard Ruby on Rails / Postgresql web app. It works quite well. It is also pretty straightforward to implement a such app in a Java/PHP MVC framework.

Let's say I would like to implement UI using React/Redux. How should I start? For example the app has calendar month view, for each day there are 20 sales calls. So the month view has 400 sales calls and clients data displayed (date, time, client name, target group).

Do I have to put 400 sales calls and 400 clients data to a Redux store to display the calendar month view? What about client data search results and pagination? In just few clicks a user can display hundreds of clients records (thousands in case of results map view). Do they belong to a Redux store? If a user modifies one sales call record, how it is persisted in central DB? What about edge cases where some uniqueness conditions have to be checked on central DB level?

Rails covers all things needed to implement my medium CRM app. When I read Redux TO DO tutorials I have a filling that they cover just 10% of what is needed to implement a full CRM app. Could you please direct me to Redux examples / tutorials how to implement a full enterprise database app (SugarCRM scale).

PS. to down voters, please write a few words what is wrong with my questions so I can learn what is appropriate to post on HN

I upvoted you, because I think this is a total legitimate question and I don't get the downvotes either. This was exactly the question I had, when I read about Redux at first. After all, the Redux docs say that "don't worry about performance, unless you have tens of thousands of objects".

So, the answer I can give you as someone, who starts to get Redux, but I'm nowhere near being an advanced user:

In your Rails app, you wouldn't render all 500k sales records on the same page as well. You'd probably have some sort of pagination. With Redux, it's the same. You don't have 500K objects in store, but only the 30 sales calls you're currently displaying. Redux contains the state of the current view and if you want to load more data, you'd have some async methods to fetch data from the DB or an API and refresh the Redux store.

Your Redux store isn't:

    [sales_call01, sales_call02, ..., sales_call500000]
It's more like:

    { viewCalls: [sales_call452, sales_call453, sales_call454] }
But still: Many tutorials are way too simple and gloss over the details. They make it sound as if Redux is the architecture and the store holds all data for the entire app.

What I haven't figured out: Say, you're using Redux on the server and need to update all 500,000 sales calls. With Redux, you'd have to write a reducer that modifies the state tree. But if 500,000 sales calls don't fit into memory, how do you do it?

it doesn't sound like redux is the right tool for the job. You're looking for a stream/file.

You could still have a redux store, but it's role would be to manage the stream(process updates, keep track of errors, kill the stream if there are too many errors) rather than store the data for the job.

Neither React nor Redux really prescribe an architecture so building an app with these libraries requires some extra thought over just using a framework. An approach I often see is to use Redux as a single state container in a monolithic flux architecture (which seems to be referred to as "Redux") but this has advantages and disadvantages, and is certainly not the only way to use these libraries.

If you're going to go down the flux path, I would suggest you map out the minimal state which can completely represent your user interface (store structure), then list all the state mutations (actions) which you expect to affect this state. That might help you build up clear picture for how to develop your CRM app.

It's good to keep in mind that at the end of the day React and Redux are just tools, and it is up to you how to use them.

My initial answer to that is that none of those questions are specific to React or Redux, but rather are common to _any_ Single Page Application, whether it be written in React+Redux, Angular 2, Angular 1, Ember, Backbone, or good ol' VanillaJS/jQuery.

More specifically: yes, you would typically make AJAX calls on app startup to load some subset of records from the server, and continue to make more AJAX calls to request more records as needed (like when the user hits the "Next Page" button). Records requested from the server would indeed be cached client-side in the Redux store. When the user modifies an entry and hits "Save", you'd make another AJAX call to persist the updated values to the backend. Again, none of that is actually Redux-specific - it's just general workflow for an SPA.

Some resources that may help. As mentioned elsewhere in the thread, I have a list of React/Redux tutorials and articles [0]. The "Redux Tutorials#Project-Based Tutorials" section [1] specifically lists tutorials that try to build something (preferably bigger than a TodoMVC app), rather than just explain the basics. In particular, the series "Building a Simple CRUD App with React and Redux" [2] is an excellent 8-part walkthrough that covers many key real-world concepts. I'll also put in a plug for my own "Practical Redux" tutorial series [3], which demonstrates some specific useful Redux techniques. Finally, you might want to glance at the "Redux Architecture and Best Practices" page [4] in my list for more articles on real-world approaches and concepts.

[0] https://github.com/markerikson/react-redux-links

[1] https://github.com/markerikson/react-redux-links/blob/master...

[2] http://www.thegreatcodeadventure.com/building-a-simple-crud-...

[3] http://blog.isquaredsoftware.com/2016/11/practical-redux-par...

[4] https://github.com/markerikson/react-redux-links/blob/master...

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact