I have a social media website where users can like/dislike posts. I have a reducer for the user and the posts. My group mate implemented the like/dislike feature as follows: whenever the user likes a post, the client calls my API with ENDPOINT 'posts/:id/like', which will increment the like count for the post in the Post Database and add the post to the user's liked posts array in the User Database. The API will return a response that contains the updated user and post, which will result in the following actions.

dispatch({
            type: LIKE_POST_SUCCESS,
            payload: res.payload.post
        })

        dispatch({
            type: UPDATE_USER_SUCCESS,
            payload: res.payload.user
        })

And the following reducers

case LIKE_POST_SUCCESS:
            return {
                ...state,
                ...action.payload
            }
case UPDATE_USER_SUCCESS:
            return {
                ...state,
                ...action.payload
            }

This approach works, but I feel his approach seems hackish in that

1. The API request is also returning the updated user alongside the updated post even though the ENDPOINT is used for the posts database.
2. He's dispatching "UPDATE_USER_SUCCESS" to update the user object in the redux store even though update user API isn't called.

Would the correct approach be

1. Calling the Like Post request
2. Updating the post store using the payload from the Like Post Request
3. Perform a "GET" request to the user database
4. Update the user store using the payload from the "GET" request