A Meteor package that creates reactive DataTables in an efficient way, allowing you to display the contents of enormous collections without impacting app performance.

• Fast: Uses an intelligent automatic data subscription so that table data is not loaded until it's needed.
• Reactive: As your collection data changes, so does your table. You can also reactively update the query selector if you provide your own filter buttons outside of the table.
• Customizable: Anything you can do with the DataTables library is supported, and you can provide your own publish function to build custom tables or tables than join data from two collections.
• Hot Code Push Ready: Remains on the same data page after a hot code push.

Although this appears similar to the jquery-datatables Meteor package, there are actually many differences:

• This package is updated to work with Meteor 1.0+.
• This package has a much smaller codebase and includes less of the DataTables library.
• This package allows you to specify a Spacebars template as a cell's content.
• This package handles the reactive table updates in a different way.
• This package is designed to work with Twitter Bootstrap 3

$ meteor add aldeed:tabular

Define your table in common code:

TabularTables = {};

Meteor.isClient && Template.registerHelper('TabularTables', TabularTables);

TabularTables.Books = new Tabular.Table({
  name: "BookList",
  collection: Books,
  columns: [
    {data: "title", title: "Title"},
    {data: "author", title: "Author"},
    {data: "copies", title: "Copies Available"},
    {
      data: "lastCheckedOut",
      title: "Last Checkout",
      render: function (val, type, doc) {
        if (val instanceof Date) {
          return moment(val).calendar();
        } else {
          return "Never";
        }
      }
    },
    {data: "summary", title: "Summary"},
    {
      tmpl: Meteor.isClient && Template.bookCheckOutCell
    }
  ]
});

And then reference in one of your templates where you want it to appear:

{{> tabular table=TabularTables.Books class="table table-striped table-bordered table-condensed"}}

Add a Mongo-style selector to your tabular component for a table that displays only one part of a collection:

{{> tabular table=TabularTables.Books selector=selector class="table table-striped table-bordered table-condensed"}}

Template.myTemplate.helpers({
  selector: function () {
    return {author: "Agatha Christie"}; // this could be pulled from a Session var or something that is reactive
  }
});

If you want to limit what is published to the client for security reasons you can provide a selector in the constructor which will be used by the publications. Selectors provided this way will be combined with selectors provided to the template using an AND relationship. Both selectors may query on the same fields if necessary.

TabularTables.Books = new Tabular.Table({
  // other properties...
  selector: function( userId ) {
    return { documentOwner: userId }
  }
});

The DataTables documentation lists a huge variety of available table options and callbacks. You may add any of these to your Tabular.Table constructor options and they will be used as options when constructing the DataTable.

Example:

TabularTables.Books = new Tabular.Table({
  // other properties...
  createdRow: function( row, data, dataIndex ) {
    // set row class based on row data
  }
});

You might have noticed this column definition in the example:

{
  tmpl: Meteor.isClient && Template.bookCheckOutCell
}

This is not part of the DataTables API. It's a special feature of this package. By passing a Spacebars Template object, that template will be rendered in the table cell. You can include a button and/or use helpers and events. In your template and helpers, this is set to the document for the current row.

Note: The Meteor.isClient && is there because tables must be defined in common code, which runs on the server and client. But the Template object is not defined in server code, so we need to prevent errors by setting tmpl only on the client.

The tmpl option can be used with or without the data option.

Here's an example of how you might do the bookCheckOutCell template:

HTML:

<template name="bookCheckOutCell">
  <button type="button" class="btn btn-xs check-out">Check Out</button>
</template>

Client JavaScript:

Template.bookCheckOutCell.events({
  'click .check-out': function () {
    addBookToCheckoutCart(this._id);
  }
});

If your table includes the global search/filter field, it will work and will update results in a manner that remains fast even with large collections. By default, all columns are searched if they can be. If you don't want a column to be searched, add the searchable: false option on that column.

If your table has a selector that already limits the results, the search happens within the selector results (i.e., your selector and the search selector are merged with an AND relationship).

The DataTables library supports calling functions on the row data by appending your data string with (). This can be used along with the dburles:collection-helpers package (or your own collection transform). For example:

Relevant part of your table definition:

columns: [
  {data: "fullName()", title: "Full Name"},
]

A collection helper you've defined in client or common code:

People.helpers({
  fullName: function () {
    return this.firstName + ' ' + this.lastName;
  }
});

Note that for this to work properly, you must ensure that the firstName and lastName fields are published. If they're included as the data for other columns, then there is no problem. If not, you can use the extraFields option or your own custom publish function.

If your table's templates or helper functions require fields that are not included in the data, you can tell Tabular to publish these fields by including them in the extraFields array option:

TabularTables.People = new Tabular.Table({
  // other properties...
  extraFields: ['firstName', 'lastName']
});

You can optionally provide an allow and/or allowFields function to control which clients can get the published data. These are used by the built-in publications on the server only.

TabularTables.Books = new Tabular.Table({
  // other properties...
  allow: function (userId) {
    return false; // don't allow this person to subscribe to the data
  },
  allowFields: function (userId, fields) {
    return false; // don't allow this person to subscribe to the data
  }
});

Note: Every time the table data changes, you can expect allow to be called 1 or 2 times and allowFields to be called 0 or 1 times. If the table uses your own custom publish function, then allow will be called 1 time and allowFields will never be called.

If you need to be sure that certain fields are never published or if different users can access different fields, use allowFields. Otherwise just use allow.

By default, a normal Meteor.subscribe is used for the current page's table data. This subscription is stopped and a new one replaces it whenever you switch pages. This means that if your table shows 10 results per page, your client collection will have 10 documents in it on page 1. When you switch to page 2, your client collection will still have only 10 documents in it, but they will be the next 10.

If you want to override this behavior such that documents displayed in the table remain cached on the client for some time, you can add the meteorhacks:subs-manager package to your app and set the sub option on your Tabular.Table. This can make the table a bit faster and reduce unnecessary subscription traffic, but may not be a good idea if the data is extremely sensitive.

TabularTables.Books = new Tabular.Table({
  // other properties...
  sub: new SubsManager()
});

Currently there is only one hook provided: onUnload

This package takes care of publication and subscription for you using two built-in publications. The first publication determines the list of document _ids that are needed by the table. This is a complex publication and there should be no need to override it. The second publication publishes the actual documents with those _ids.

The most common reason to override the second publication with your own custom one is to publish documents from related collections at the same time.

To tell Tabular to use your custom publish function, pass the publication name as the pub option. Your function:

• must accept and check three arguments: tableName, ids, and fields
• must publish all the documents where _id is in the ids array.
• should publish only the fields listed in the fields object, if one is provided.
• should do any necessary security checks
• may also publish other data necessary for your table

Suppose we want a table of feedback submitted by users, which is stored in an AppFeedback collection, but we also want to display the email address of the user in the table. We'll use a custom publish function along with the reywood:publish-composite package to do this. Also, we'll limit it to admins.

server/publish.js

Meteor.publishComposite("tabular_AppFeedback", function (tableName, ids, fields) {
  check(tableName, String);
  check(ids, Array);
  check(fields, Match.Optional(Object));

  this.unblock(); // requires meteorhacks:unblock package

  return {
    find: function () {
      this.unblock(); // requires meteorhacks:unblock package

      // check for admin role with alanning:roles package
      if (!Roles.userIsInRole(this.userId, 'admin')) {
        return [];
      }

      return AppFeedback.find({_id: {$in: ids}}, {fields: fields});
    },
    children: [
      {
        find: function(feedback) {
          this.unblock(); // requires meteorhacks:unblock package
          // Publish the related user
          return Meteor.users.find({_id: feedback.userId}, {limit: 1, fields: {emails: 1}, sort: {_id: 1}});
        }
      }
    ]
  };
});

common/helpers.js

// Define an email helper on AppFeedback documents using dburles:collection-helpers package.
// We'll reference this in our table columns with "email()"
AppFeedback.helpers({
  email: function () {
    var user = Meteor.users.findOne({_id: this.userId});
    return user && user.emails[0].address;
  }
});

common/tables.js

TabularTables.AppFeedback = new Tabular.Table({
  name: "AppFeedback",
  collection: AppFeedback,
  pub: "tabular_AppFeedback",
  allow: function (userId) {
    // check for admin role with alanning:roles package
    return Roles.userIsInRole(userId, 'admin');
  },
  order: [[0, "desc"]],
  columns: [
    {data: "date", title: "Date"},
    {data: "email()", title: "Email"},
    {data: "feedback", title: "Feedback"},
    {
      tmpl: Meteor.isClient && Template.appFeedbackCellDelete
    }
  ]
});

Some useful tips

Template.myTemplate.events({
  'click tbody > tr': function (event) {
    var dataTable = $(event.target).closest('table').DataTable();
    var rowData = dataTable.row(event.currentTarget).data();
  }
});