This example shows how to use the DataTable's `recordType` attribute to create columns whose values are calculated from other columns. It was contributed by community member Todd Smith.
In the table below, the "Market Value" and "Gain or Loss" columns are created from data in the other columns by configuring them as attributes with `getter`s in the table's `recordType` configuration.
DataTable's `recordType` attribute is used to specify the class used to store the record instances in your table. If you don't configure your DataTable's `recordType`, a `Y.Model` subclass will be created for you based on the keys in the data that you fill the table with.
As well as accepting Model subclasses, `recordType` can be passed an object that corresponds to a class's `ATTRS` collection. From this, DataTable will create the Model subclass accordingly. This is useful if you want to provide default values, or validate or transform incoming values (such as turning numeric strings into numbers).
``` var table = new Y.DataTable({ columns: ['cost', 'price', 'profit'], recordType: { cost: { value: 0.0001, // default value validator: function (val) { return Y.Lang.isNumber(val) && val > 0; } }, price: { setter: function (val) { val = +val; // coerce numeric strings to numbers return isFinite(val) ? val : Y.Attribute.INVALID_VALUE; } } }, ... }); ```Another thing you can configure with attributes is a `getter`, which is responsible for returning a value to `instance.get('yourAttribute')` calls. Since DataTable uses this method to get cell contents for columns, you can use `getter`s to create columns whose content is calculated on the fly.
``` var table = new Y.DataTable({ columns: ['cost', 'price', 'profit'], recordType: { cost: {}, price: {}, profit: { getter: function () { return this.get('price') - this.get('cost'); }, readOnly: true } }, ... }); ```The `data` populating the table need only include `cost` and `price`. The `profit` column will be populated automatically based on the other two.
It is possible to create calculated columns using column `formatter`s as well, but if you need that column to be sortable, you need to associate that column with a single `key`, which may be insufficient, or configure the column with a `sortFn`.
Using an attribute with a `getter`, you are creating a new field with a `key` that corresponds to the full values you want to sort by using the default sort mechanism. And you can even add `formatter`s on top of the raw, sortable, calculated value.