Right now, there is no method to dispose of a query (much like disposing a group in crossfilter), other than clearing the column definition (which disposes the dimension and all of it's groupings).

There should be a .clear() function returned with the .query() promise that when called will clear the query/group from memory and deregister all listeners.

I haven't had the time to isolate this one into a test but if I create a filter using a function:

const fn => () { /* some function /* }
universe.filter('search', fn)

Then later try to add another filter I get the error: f is not a function at filterIndexFunction.

I have determined this is because the filter function gets converted to a string here: https://github.com/crossfilter/universe/blob/master/src/filters.js#L38

My suggested solution is:

      var newFilters = _.clone(service.filters, true)
      _.map(newFilters, function(fil, i) {
        if (fil.type === 'function') {
          fil.function = service.filters[i].function
          fil.value = service.filters[i].value
        }
      })

I got this error after installing it. Using it with Webpack and Babel. I don't know if it helps but I was using Crossfilter.js without issues.

Let me know if you need more information.

Hi,
I really appreciate the effort of lowering the barrier of entry to crossfilter - in particular providing some kind of grammar for querying multivariate dataset. However, playing around the library is not extremely straightforward at the moment, and I found that its added value over defining your own dimensions and groups in crossfilter directly was not obvious from the provided code samples.

Maybe porting the Airline on-time performance example to universe could help a user like myself see/understand/delve into a more comprehensive usecase (e.g. filtering the dataset on multiple dimensions, see how we can add listeners to filtering actions, ... ).

I actually tried to port this example myself - without success (at least given the timeframe I allowed myself to test out the lib - and the available amount of brain cells...).

Thanks anyway,
C.

P.S. There are some helpful hints/example in the gitter channel though - saw them a bit too late!
P.P.S. Oh - this should be highlighted ! http://hypercubed.github.io/Project-Chi/#/examples/universe

i'm working on a project that need capabilities that universe library have, but i had some trouble implementing the squash method.
This is my code, basically its the same with the example provided in the readme section

universe([
    {date: "2011-11-14T16:17:54Z", quantity: 2, total: 190, tip: 100, type: "tab", productIDs: ["001"]},
    {date: "2011-11-14T16:20:19Z", quantity: 2, total: 190, tip: 100, type: "tab", productIDs: ["001", "005"]},
    {date: "2011-11-14T16:28:54Z", quantity: 1, total: 300, tip: 200, type: "visa", productIDs: ["004", "005"]},
    {date: "2011-11-14T16:28:54Z", quantity: 3, total: 285, tip: 100, type: "tab", productIDs: ["001", "002", "005"]}
]).then(function(myUniverse){
    myUniverse.query({
        groupBy: 'type', // GroupBy the type key
        select: {
            $count: true, // Count the number of records
            quantity: { // Create a custom 'quantity' column
                $sum: 'quantity' // Sum the quantity column
            },
        },
        // Limit selection to rows where quantity is greater than 50
        filter: {
            tip: {
                $gt: 50
            }
        },
    }).then(function(result) {
          result.squash(null, null, {
            count: '$sum',
            quantity: {
                sum: '$sum'
            }
        },'Total').then(function(resultq) {
                   console.log(JSON.stringify(resultq.data));
            })
    })

universe_squash.txt

and this is the output
[{"key":"Total","value":{"count":4,"quantity":{"0":"sum"}}}]

what i was expecting from what i read in the example is something like the output below
[{"key":"Total","value":{"count":4,"quantity":{"sum":8}}}]

Am i calling the squash method the wrong way or is this a bug?
Sorry for my bad English.

If an aggregation is the only nested item in an alias, then the resulting data should be flatten into that alias, eg.

.query({
  groupBy: 'type',
  select: {
    myColumn: {
      $sum: 'total'
    }
  }
})

should result in:

[{
  key: 'type1',
  value: {
    myColumn: 30
  }
}, {
  key: 'type2',
  value: {
    myColumn: 15
  }
}, {
  key: 'type3',
  value: {
    myColumn: 24
  }
}, {
  key: 'type4',
  value: {
    myColumn: 12
  }
}]

this should also work with the string syntax. eg.

.query({
  groupBy: 'type',
  select: {
    myColumn: '$sum(total)'
  }
})

Right now, dimensions are tracked and managed with easy references in universe.columns, but the queries made on those columns are not tracked anywhere.

universe.columns[n].queries would be a great place to put these query objects.

We could store there:

• the original query object
• crossfilter group
• listeners/callbacks
• post-aggregations (in the future)
• etc.

the result object in the console will not contain a crossfilter dimension as stated in the documentation.

universe([
  {date: "2011-11-14T16:17:54Z", quantity: 2, total: 190, tip: 100, type: "tab", productIDs:["001"]},
  {date: "2011-11-14T16:20:19Z", quantity: 2, total: 190, tip: 100, type: "tab", productIDs:["001", "005"]},
  {date: "2011-11-14T16:28:54Z", quantity: 1, total: 300, tip: 200, type: "visa", productIDs:["004" ,"005"]},
  {date: "2011-11-14T16:30:43Z", quantity: 2, total: 90, tip: 0, type: "tab", productIDs:["001", "002"]},
  {date: "2011-11-14T16:48:46Z", quantity: 2, total: 90, tip: 0, type: "tab", productIDs:["005"]},
  {date: "2011-11-14T16:53:41Z", quantity: 2, total: 90, tip: 0, type: "tab", productIDs:["001", "004" ,"005"]},
  {date: "2011-11-14T16:54:06Z", quantity: 1, total: 100, tip: 0, type: "cash", productIDs:["001", "002", "003", "004" ,"005"]},
  {date: "2011-11-14T16:58:03Z", quantity: 2, total: 90, tip: 0, type: "tab", productIDs:["001"]},
  {date: "2011-11-14T17:07:21Z", quantity: 2, total: 90, tip: 0, type: "tab", productIDs:["004" ,"005"]},
  {date: "2011-11-14T17:22:59Z", quantity: 2, total: 90, tip: 0, type: "tab", productIDs:["001", "002", "004" ,"005"]},
  {date: "2011-11-14T17:25:45Z", quantity: 2, total: 200, tip: 0, type: "cash", productIDs:["002"]},
  {date: "2011-11-14T17:29:52Z", quantity: 1, total: 200, tip: 100, type: "visa", productIDs:["004"]}
])
.then(function(myUniverse){
  return myUniverse.query({
    groupBy: 'productIDs',
    select: {
      $count: true,
    }
  })
  .then(function(res){
    console.info('res:',res)
   })
})

I've noticed that the code contains mixed styles (ASI and semicolons). Would you consider a PR to add a linter? I prefer XO. Standard and semi-standard are both good as well. Personally, I prefer semicolons and spaces... but I will set up the linter however you prefer.

Could we have a new tag version including the latest code from master ?

I'd personally need that to be able to integrate the latest fix into bower

The documentation shows usage of a filterAll method on a universe instance (https://github.com/crossfilter/universe#explore-your-data). However this method, while it exists in the ./src/filters.js file (https://github.com/crossfilter/universe/blob/master/src/filters.js#L48) ), is not exposed on the universe instance. Furthermore, the filterAll method as it exists is not sufficient.

Here is a gist of a mocha tests for filterAll: https://gist.github.com/Hypercubed/fb7eac133e8675f3e193de36cd3d3459

You will see three tests:

1. filterAll method is not exposed
2. filterAll method clears the universe.filters object but does not clear the individual filters.
3. A promise chain that does clear each filter and clears the filter object.

Given this I think the proper filterAll method should be:

function filterAll() {
  return Promise.all(_.keys(service.filters).map(function(key) {  // clearing each filter
    return service.filter(key)
  }))
  .then(function() {  // clearing filters object
    return applyFilters({})
  })
}

or equivalent.

Currently, you are using gulp and browserify to build universe.js and universe.min.js files in the root. package.json uses src/universe.js as main. The tests use '../universe.js' (the browserified version). Can I suggest that the tests use src/universe.js and we .gitignore the built files? Are you planning to support bower?

Also, as an aside, I mostly switched from using gulp to using only npm scripts. Since this build is fairly simple we might consider the same here.

On chrome v51 http://codepen.io/tannerlinsley/pen/oxjyvg?editors=0010