GroupBy

GroupBy is used to group the rows that have the same values in one or more columns. For example, if a particular column has the same values in different rows, groupBy will arrange these rows in a group.

When you apply groupBy into a DataTable, it transforms the data and creates a new DataTable. The syntax to apply groupBy to the DataTable in the DataStore is given below:

FusionCharts.DataStore.Operators.groupBy(groupConfigArr, aggrConfigArr);

In the above code:

  • FusionCharts.DataStore.Operators is the namespace in which groupBy method resides.
  • groupBy is the method applied
  • groupConfigArr and aggrConfigArr are the arguments

GroupBy method takes two arguments which are:

  • An array of the group by configurations.

  • An array of the aggregation configurations.

Groupby Configuration

In the first argument of the groupBy method, each group of configuration object can have the following attributes:

column

Name of the column from the DataTable created using schema.

outputAs

outputAs is an attribute used to rename the column on which groupBy has been applied. The renaming of the column reflects on the new dataTable created after groupBy.

If outputAs is not applied, the column name remains the same as the previous DataTable.

timeUnit

This attribute is only applicable if you group on date/time. This attribute specifies how the date column should be grouped. For example, if you set timeUnit to Month, the grouping will be applied month wise. The month is one of the DatetimeUnits enum and not any random string.

DateTimeUnits is a separate enum of FusionCharts.Utils. DateTimeUnits consists of:

  • Year
  • Quarter
  • Month
  • Week
  • Day
  • Hour
  • Minute
  • Second
  • Millisecond

It is mandatory to set the unit using timeUnit attribute at the time of date/time grouping.

Refer to the code below:

column: 'OrderDate',
timeUnit: FusionCharts.Utils.DatetimeUnits.Month

outputFormat

outputFormat attribute is used to set the format of the column being grouped. To set the date/time format of the chart, set the value of outputFormat attribute as per your requirement.

If the outputFormat of the data is not specified, an example of a default date/time format is given below:

unitDefaultFormats = {
    'Year': '%Y', // 2018
    'Month': '%b %Y', // Jan 2018
    'Day': '%b %d, %Y', // Jan 01, 2018
    'Hour': '%b %d, %Y %H hrs', // Jan 01, 2018 23 hrs
    'Minute': '%b %d, %Y %H:%M', // Jan 01, 2018 23:34
    'Second': '%b %d, %Y %H:%M:%S', // Jan 01, 2018 23:34:26
    'Millisecond': '%b %d, %Y %H:%M:%S:%L' // Jan 01, 2018 23:34:26:123
};

The detailed table specifing the date/time format, is given here.

startValue

startValue attribute specifies the starting point of the binning. If you don't apply this attribute while grouping, the lowest available value is used as the starting point.

startValue attribute is applicable only when you are applying grouping operation to date/time column.

weekStartFrom

weekStartFrom attributes specifies the weekday from where you want to start the week at the time of grouping. This attribute is applicable only when weekly binning is applied at the time of grouping.

Weekdays is a separate enum of FusionCharts.Utils. Weekdays consists of:

  • Sunday
  • Monday
  • Tuesday
  • Wednesday
  • Thursday
  • Friday
  • Saturday

binSize

As the name suggests, this attribute sets the size of the bin. The default value of this attribute is 1.

binSize attribute is applicable only when you are applying grouping operation to date/time column.

For example:

[{
    column: 'Import date',
    timeUnit: FusionCharts.Utils.DatetimeUnits.Week,
    weekStartsFrom: FusionCharts.Utils.Weekdays.Monday,
    binSize: 2,
    startValue: +new Date(2018/1/15),
    outputAs: 'Fortinightly sum of imports'
}],

In the above code, binSize attribute has been set to 2, which means every bin has 2 weeks data (as the timeUnit has been set to Week). The bin calculation starts from startValue which has been set to a particular date.

Aggregation Configuration

In the second argument of the groupBy method, you can apply the following attributes to each group of configuration:

column

Name of the column from theDataTablecreated using schema..

operation

operation is an attribute which denotes the aggregation method to be applied on each created group. It accepts string value. The list of values accepted by this attribute are:

  • sum
  • avg
  • count
  • min
  • max
  • first
  • last
  • variance
  • stddev

outputAs

outputAs is an attribute which is used to rename the column, on which groupBy has been applied. The renaming of the column reflects in the new dataTable created after groupBy.

If the outputAs is not applied, the column name will be previous column name-operation.

Now, lets apply groupBy operation to the DataTable given below:

Order Date Country Sales Quantity Shipping Cost
1/22/2011 Australia 59.724 6 27.43
1/22/2011 United States 125.248 3 3.64
1/22/2011 Australia 6.318 1 1.77
1/24/2011 Australia 110.808 3 9.92
1/24/2011 United States 40.08 6 4.31
1/24/2011 India 59.25 5 4.27
1/24/2011 United States 5.94 3 0.95
1/26/2011 India 79.38 3 13.82
1/26/2011 India 342.51 7 13.22

The data structure to group the data is shown below:

var groupQuery = groupBy(
    [{
        column: 'Order Date', // Column to which groupBy has been applied
        timeUnit:FusionCharts.Utils.DatetimeUnits.Day
    }],
    [{
        column: 'Sales',
        operation: 'sum',
        outputAs: 'Total_Sales'
    }]
);

The new DataTable after running the above code looks like:

Order Date Total_Sales
1/22/2011 191.29
1/24/2011 216.078
1/26/2011 421.89