React.js和Redux-如何创建一个“流行”页面

Question

I'm trying to create a page on my app that shows a list of the most popular events to go to. The events that have the most 'interested in' and 'going to' results.

I'm using knex and sqlite3 for my database and currently have junction tables for users and the events they're interested in, and another for the user and events they're going to.

How would I structure a page to return the most interacted with events? Do I need to add some sort of count column to the events table and return events with the biggest count (having a function that each time it is selected the count goes up...)?

I'm mainly not sure about where to start with this/how to approach this.

Thanks!

---

This is how my data is structured so far:

Events Table:

exports.up = (knex, Promise) => {
  return knex.schema.hasTable('events').then(function (exists) {
    if (!exists) {
      return knex.schema.createTable('events', (table) => {
        table.increments('id').primary()
        table.string('eventName')
        table.date('date_start')
        table.string('time_start')
        table.date('date_end')
        table.string('time_end')
        table.string('image')
        table.string('description')
        table.string('access')
        table.string('tickets')
        table.string('restrictions')
        table.string('eventType')
      })
    }
  })
}

exports.down = (knex, Promise) => {
  return knex.schema.dropTable('events')
}

Seed Data: user_goingto_events

exports.seed = function(knex, Promise) {
  // Deletes ALL existing entries
  return knex('user_goingto_events').del()
    .then(function () {
      // Inserts seed entries
      return knex('user_goingto_events').insert([
        {id: 1, user_id: 1, event_id: 1},
        {id: 2, user_id: 1, event_id: 2},
        {id: 3, user_id: 1, event_id: 3},
        {id: 4, user_id: 2, event_id: 4},
        {id: 5, user_id: 2, event_id: 5},
        {id: 6, user_id: 2, event_id: 6}
      ])
    })
}

Seed Data: user_interestedin_events

exports.seed = function(knex, Promise) {
  // Deletes ALL existing entries
  return knex('user_interestedin_events').del()
    .then(function () {
      // Inserts seed entries
      return knex('user_interestedin_events').insert([
        {id: 1, user_id: 1, event_id: 4},
        {id: 2, user_id: 1, event_id: 5},
        {id: 3, user_id: 1, event_id: 6},
        {id: 4, user_id: 2, event_id: 1},
        {id: 5, user_id: 2, event_id: 2},
        {id: 6, user_id: 2, event_id: 3}
      ])
    })
}

Answer 1

The approach I would take would be to SUM() the event counts for each of the 2 tables, and then UNION them together:

In SQL that looks like:

select event_id, count(*) as cnt_interest, 0 as cnt_going
   from user_interestedin_events
   group by event_id
UNION
select event_id, 0 as cnt_interest, count(*) as cnt_going
   from user_goingto_events
   group by event_id;

Then I would use that result as the input to summarize the counts producing a single row per event. Theoretically, if you had a result cursor named 'part1data', then the SQL would look like this:

select eventName, event_id, sum(cnt_interest) as sum_interest, sum(cnt_going) as sum_going
   from part1data
   left join events on events.id = event_id
   group by event_id;
   order by sum(cnt_interest)+sum(cnt_going) desc

You can see I ordered it by popularity, and also included the event name via a left join for the users convenience.

Now to put that in Knex. Since this is a bit complicated, I would just embed it as a raw() SQL and run it as follows:

const rawRpt = 
 `select eventName, event_id, sum(cnt_interest) as sum_interest, sum(cnt_going) as sum_going
   from 
     ( select event_id, count(*) as cnt_interest, 0 as cnt_going
          from user_interestedin_events
          group by event_id
        UNION
       select event_id, 0 as cnt_interest, count(*) as cnt_going
          from user_goingto_events
          group by event_id
     )
   left join events on events.id = event_id
   group by event_id;
   order by sum(cnt_interest)+sum(cnt_going) desc`

return Knex.raw(rawRpt);
   .on('query-error', function(ex, obj) {
        console.log("KNEX query-error ex:", ex, "obj:", obj);
   })
   .then(function(retList) {
        console.log("Returns:", retList);
   });

You can probably see that I embedded the first query inside the second. I would also note, for 90% of the Knex SQLs that I do, I do NOT use raw(). But in this case, the complexity, the creation of multiple calculated fields, and having no input variables ... this is just easier.

I hope this helps. (Disclaimer - I didn't run the code so it's probably not perfect.)