{{ transformations:GroupAction.png}}
====== GROUP ======
Category: Transform / Advanced\\

\\ 
=====Description=====
Replace multiple values with a single value (constant or calculated). \\

\\ 
=====Use cases=====
  * Creating flags that can later be used in expressions to handle the values within the group (set).  This can be more convenient than using lengthy if..then..else statements to group a set of values.
  * Group values for charts (e.g. changing insignificant dimension values to "Other")
  * Calculating new values of a column only for particular values in another column, leaving the rest unchanged.
  * Create secondary demographic sets, combining specified values into rollup groups.

\\ 
=====Action settings=====
^Setting^Description^
|Create/Replace column|Select whether the results will replace the values in an existing column or be created in a new column.\\  Options: //Create new column// (and provide a column name), or //Replace existing// (and select an existing column).|
|Group (flag) as<sup>*</sup>|Provide a static value or expression that will be used as the flag value.  All selected values will be converted\\ to this value. Options:  //It's an expression// or //It's a text or number//. |
|Lookup column|Select the column with the values to be grouped.  Once chosen, a list of values present in the column will be\\ displayed at the bottom of the action settings pane. |
|Grouping|Choose whether the selected values will be replaced with the group (flag) value, or whether all values //except//\\ the selected values will be replaced.  Options: //Group selected//, or //Group all but selected//.|
|Value selection|Select the values to be replaced, or excluded from replacement, in the list.|
<sup>*</sup> Setting can be specified using a [[:parameters|parameter]].

\\ 
=====Remarks=====
Values that are not converted to the group (flag) value retain their original value when using the //Group selected// option and selecting values.

Use the //Group selected// option, and select values to evaluate when dealing with an expected set of incoming values (see "Status" column in Example #1, below).  When dealing with a column containing varying values (e.g. calculated values), use the "Group all but selected" option and do not select any values to evaluate any/all values that pass into this action (see Example #2, below).

When using an expression for the //Group (flag) as// setting in conjunction with the //Group selected// option (and selected values), values that are not flagged by the expression retain their original values.  However, when using an expression in conjunction with the //Group all but selected// option (with no selections), the expression determines the output (see Example #2 below).

\\ 
=====Examples=====

====Example #1====
>Flag accounts that are past due.

===Before (source table)===
^Customer ^Status ^
|Alice  |Overdue  |
|Pierre  |Paid  |
|Iryna|Past due  |
|Michael  |Pending  |

===After (result table)===
^Customer ^Status ^Late flag  ^
|Alice  |Overdue  |**Late**  |
|Pierre  |Paid  |Paid  |
|Iryna|Past due  |**Late**  |
|Michael  |Pending  |Pending  |

===Action parameters===
>Creaate new column: Late flag
>Group (flag) as:  Late
>Group selected (Overdue, Past due)

\\ 
====Example #2====
>Flag transactions that exceed $5000 as "High Risk" for further review.

===Before (source table)===
^Transaction ID ^Customer ^Amount ($) ^
|T001  |John  |  3200|
|T002  |Emma  |  7500|
|T003  |Shira  |  4200|
|T004  |Michael |  5500|

===After (result table)===
^Transaction ID ^Customer ^Amount ($) ^Review ^
|T001  |John  |  3200|  |
|T002  |Emma  |  7500|**High Risk**  |
|T003  |Shira  |  4200|  |
|T004  |Michael |  5500|**High Risk**  |

===Action parameters===
>Create new column: Review
>Group (flag) as:  (expression) if [Amount ($)] >=5000 then "High Risk"
>Group all but selected (//None// selected)

\\
=====See also=====
  * [[transformations:addcolumn|Calculate new column(s)]]
  * [[transformations:replacecolumns|Modify column(s)]]
  * [[syntax:functions|Functions:  In()]]
  * [[syntax:functions|Functions:  Match()]]