Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: Migrated to Confluence 5.3

GROUP BY

GROUP BY allows users to apply other aggregate functions to sub-sets of the set of entities matching the query.

This is the fundamental building block for analysing information by Type, Status, Priority etc.

Syntax

Code Block
languagenone
GROUP BY <FieldSelector> [AS "Alias"] { <AggregationFunction1>, <AggregationFunction2> ... } AS "Alias"

The "FieldSelector" can currently be either a field name e.g.

Code Block
languagenone
GROUP BY Status { COUNT }

Which would display results like this:

Or can be a function call:

Code Block
languagenone
GROUP BY Fill(Status) { COUNT }

Which in this case would fill out the results to include all possible values (not just values included in the result of the query).

In the above GROUP BY statement, the function "Fill" will expand the GROUP BY to include all values from the Status picklist (which means the results will include 0-counts/0-sums).

Functions

Fill

The Fill function takes a single argument, the name of the field to GROUP BY. It will expand the results to include all groups, not just the groups which include 1 or more matches within the query.

Example:

Code Block
languagenone
GROUP BY Fill(Priority) { COUNT }

Month

The Month function takes at least one argument, the name of the field to GROUP BY. The field must be a date/time field. It will group the results based on the Month of the date supplied (all results will be grouped by the first day of the month they belong to).
Optionally it can also include a Format and/or Range function (see the Format and Range functions below).

Example:

Code Block
languagenone
GROUP BY Month(CreatedAt) { COUNT }

This above aggregation would display results grouped by the first day of each month, as shown in this screen shot:

Day

The Day function takes at least one argument, the name of the field to GROUP BY. The field must be a date/time field. It will group the results based on the Day of the date supplied (all results will be grouped by the day of the year they belong to, regardless of time). Optionally it can also include a Format and/or Range function (see the Format and Range functions below).

Example:

Code Block
languagenone
GROUP BY Day(LastUpdatedAt) { COUNT } WHERE LastUpdatedAt >= DateSubtract(Now(), "2 weeks")

Note: When executing groupy by queries, results will be confined to only those days included in the result set:

To return the empty days as well, across a date range, use the "Range" function (see below for more details):

Year

The Year function takes at least one argument, the name of the field to GROUP BY. The field must be a date/time field. It will group the results based on the Year of the date supplied (all results will be grouped by the Year they belong to). Optionally it can also include a Format and/or Range function (see the Format and Range functions below).

Example:

Code Block
languagenone
GROUP BY Year(CreatedAt) { COUNT }

Quarter

The Quarter function takes at least one argument, the name of the field to GROUP BY. The field must be a date/time field. It will group the results based on the Quarter of the year they belong to. Optionally it can also include a Format and/or Range function (see the Format and Range functions below).

Example:

Code Block
languagenone
GROUP BY Quarter(CreatedAt) { 
    GROUP BY Project {
        COUNT
    }
}

Format

The format function is a second-level function that can be passed to the various "Date" functions (Month, Day, Year & Quarter). It will be used to format the value of the date for display purposes - this allows you to convert the value in a more friendly format (so for example when grouping by Month, you might want to output the date as "June - 2011", instead of the default "2011-06-01".

The format strings correspond to the Microsoft .Net Date/Time format strings, as documented here.

Note: The resulting string must be unique from all other group values strings, so for example if grouping by day, you could not provide a format string that only includes the year, as this would result in non-unique group names.

Example:

Code Block
languagenone
GROUP BY Year(CreatedAt, Format("yyyy")) { COUNT } WHERE EntityType = Requirement

Range

By default the values will range from the natural start and end date of the results, but in some cases you may wish to extend the range further for the purposes of reporting (for example if you want to have the range include particular start and end dates).

Range takes 2 arguments, both date/time values, being the start and end dates of the range.

Example:

Code Block
languagenone
GROUP BY Day(LastUpdatedAt, Range('2011-1-1','2012-1-1')) { COUNT } WHERE Project = 'Project X' AND EntityType = Requirement

Examples

GROUP BY packages to find when the first script assignment was created, and the last script assignment was updated (to see where activity in the project is occurring).

Code Block
languagenone
GROUP BY Package { Min(CreatedAt) AS "First Activity", Max(LastUpdatedAt) AS "Last Activity", COUNT AS "Total"} AS "Package" WHERE Project = "Project X" AND EntityType = ScriptAssignment

A variation on the previous example, using the full path of the packages, as opposed to just the package name:

Code Block
languagenone
GROUP BY PackagePath { 
    Min(CreatedAt) AS "First Activity", 
    Max(LastUpdatedAt) AS "Last Activity", 
    COUNT AS "Total"
} AS "Package"
WHERE Project = "Project X" AND EntityType = ScriptAssignment

Here is how the output for the above query would look:

Find the number of entities being created per month over the last 3 years:

Code Block
languagenone
GROUP BY Month(CreatedAt, RANGE('2010-1-1', '2012-1-1'), Format('MMMM yyyy')) { COUNT }

The above example will GROUP BY date with labels like "January 2010" etc.

Below is an example of nested GROUP BY aggregations - this example gives a break down of each incident, by component, then type and finally priority, where the status is "Open" or "Reopened"

Code Block
languagenone
GROUP BY Component {
  GROUP BY Type {
    GROUP BY Priority {
      COUNT
    }
  }
}   
WHERE Status = Open

The above query would return results like this, using nested tree nodes:

Note: Though we display the above query over multiple lines (for clarity), this is entirely optional and nested GROUP BY's can all appear on a single line. Because pressing enter while typing executes the query, you need to hold the "SHIFT" key while pressing enter to move the cursor to the next line of the text editor.

See Also