Add or edit a data configuration variable

When creating or editing a data configuration, you can add or modify configuration variables as described below. You can also delete data configuration variables by clicking the Row Action menu () for the variable on the Modify Configuration page and then clicking Delete.

If you edit or delete data configuration variables after the data configuration has been used to create a data mining run or a query-based case series, you might not be able to view results for that run or that case series properly. For example, if you use a variable as an item variable in a data mining run, then change its variable type so that it is no longer an item variable, Oracle Empirica Signal will not be able to display the run results properly.

In the left navigation pane, click the Settings icon ().
In the Configure System section, click Manage Configurations.
Click the Row Action menu () for the data configuration, then click Edit.
Click the Add New Variable link, or click the Row Action menu () for a configuration variable, and then click Edit.
In the Name field, enter the name of the variable.

Variable names display throughout the application, including when users select criteria for viewing data mining results, defining queries, or creating report definitions. For example, a column might be named SYMPTOM in the Oracle table, but renamed to appear as Event in the application.

Note:
Commas and vertical bars are not permitted in the Name field.
In the Description field, type a description of the variable.
Next to the Table field, click Select Table/Column.

A data configuration can refer to one or more tables in a database, and one or more columns in each table. For example, you might create a data configuration that gets data from a demographics table, a drugs table, and an events table. However, if you already created an Oracle view that combines data from those three tables, you can refer to that view instead of to the three tables.
1. From the Available Columns list, select the column and click Save. You can select any table in the Oracle account to which you are connected, and any column in that table.
  
  The Edit Variable page appears with the column you selected displayed in the Column field.
2. In the Prefix field, enter a prefix.
  
  A prefix is required only if the variable type is Event, Drug, or Generalized Item. The alphabetical order of prefixes determines the column names for the variables in the complete results table. For example, if the prefix for the drug variable is D and the prefix for the event variable is E, the drug variable is in the ITEM1 column and the event variable is in the ITEM2 column.
  
  Note:
  Multiple variables with the same prefix cannot be used as item variables in the same data mining run.

From the Variable Type drop-down list, select a variable type.

To determine which variable type is appropriate, you might want to know the Oracle data type of the column. Select Table/Column next to the Table field. In the available Columns list, the data type appears in parentheses after the column name.

The variable type can be one of the following:

Variable Type	Meaning
Other	Identifies a variable that is not available as an item variable in data mining runs. If you want the variable to be available for subsetting or stratification of data mining runs, check Use as Subset Variable or Use as Stratification Variable. The column must have an Oracle data type of CHAR, VARCHAR2, NUMBER, FLOAT, INTEGER, DECIMAL, or SMALLINT.
Age	Identifies the patient age. The column must have an Oracle data type of CHAR, VARCHAR2, NUMBER, INTEGER, or SMALLINT. The selection type for this type of variable must be either Distinct Value List or Continuous.
Drug	Identifies the drug or vaccine. All variables of this type are available as item variables in data mining runs. The column must have an Oracle data type of CHAR or VARCHAR2. A prefix must be specified for this type of variable, and its selection type must be either Distinct Value List or Select from Drug Hierarchy.
Event	Identifies the event or symptom. All variables of this type are available as item variables in data mining runs. The column must have an Oracle data type of CHAR or VARCHAR2. A prefix must be specified for this type of variable, and its selection type must be either Distinct Value List or Select from Event Hierarchy.
External Report ID	For use when you want to include a link to an external (third-party) application from case details. (To do so, you specify an external application URL on the Edit Configuration Details page.) To use this feature, there must be a mapping of Oracle Empirica Signal case IDs to the IDs used by the external application. A suggested way to provide the mapping is to include in the demographics table an additional column, such as EXTERNAL_ID, to contain the external ID associated with each case ID in the demographics table. In the data configuration, you must include a variable for that column and assign to it the variable type External Report ID. There can be only one variable of the type External Report ID in the data configuration. If the case IDs used by the external application are the same as the Oracle Empirica Signal case IDs, an external report ID is not needed to link to the external application.
Gender	Identifies the patient gender. The column must have an Oracle data type of CHAR or VARCHAR2. The selection type for this type of variable must be either Distinct Value List or Free Text.
Generalized Item	Identifies variables other than drug or event that are available as item variables in data mining runs. For example, if the Gender variable's type is Generalized Item, users can perform data mining for combinations for Drug, Event, and Gender values. The column must have an Oracle data type of CHAR or VARCHAR2. A prefix must be specified for this type of variable, and its selection type must be Distinct Value List.
Report ID	Identifies the case ID (report ID or ISR). The column must have an Oracle data type of CHAR or VARCHAR2. You must include a Report ID variable for each source database table referenced by a configuration variable. As a result, the configuration might have several Report ID variables. The selection type for a Report ID variable must be Distinct Value List, and the variable cannot be identified as a subset variable or stratification variable. Note: In the Query Wizard, this type of variable allows entry of a string of report IDs separated by commas.
Valid Start Date, Valid End Date	Applies only if the data configuration supports timestamped data (that is, Database is Timestamped is checked for the data configuration). Identifies the column in the source data that contains a start date or end date for each record. You must include a Valid Start Date variable and a Valid End Date variable for each source database table referenced by a configuration variable. As a result, a data configuration might have several start and end date variables. The selection type for this type of variable must be Continuous.

Note:

If the variable type is Drug, you must also specify a variable subtype of Generic, Ingredient, or Trade. Each subtype can be used for only one data configuration variable. The subtype makes the variable's values available for selection on the Drug Profile page as a generic name, an ingredient, or a trade name.

If you want the variable to be available for subsetting data mining runs, check Use as subset variable. Typically, the Variable Type of a subset variable is Other. You can define multiple variables in a data configuration as subset variables, though only one subset variable can be selected for a given data mining run.
If you want the variable to be available for stratification of a data mining run, check Use as stratification variable. Typically, the variable type of a stratification variable is Other. Although you can define a variable with a variable type of Drug, Event, or Generalized Item as a stratification variable, you cannot use the same variable as both an item variable and a stratification variable in the same run.

Note:
Stratification variables that are not associated with data transformations are available as additional covariates in logistic regression runs.

From the Selection Type drop-down list, choose a selection type, which determines how users specify values for the variable when performing activities such as creating a query.

Selection Types

Selection Type	Meaning
Distinct Value List	Users choose values from a list. Variables with a variable type of Generalized Item, or of Other (with Use as subset variable or Use as stratification variable checked), must use this selection type. For variables with this selection type, you can identify a unique values table to supply the list of values. If you do not define a unique values table, the application creates one automatically when you validate the data configuration.
Select from Drug Hierarchy, Select from Event Hierarchy	Use one of these selection types only if a drug or event hierarchy account is specified at the level of the configuration as a whole and you want users to be able to browse for and select terms from the hierarchy. (Users can also select values from a list.) If you do not specify this selection type, users need to select hierarchy terms from an alphabetized list instead of the hierarchy. See About drug and event hierarchies. Note: If there is a hierarchy table specified for the individual variable, select the variable type Distinct Value List for the variable.
Continuous	Users specify a range of values by entering From and To values.
Free Text	Users type in a text string for which matches will be found.

In the Hierarchy Level field, select the hierarchy level.

If the selection type of the variable is Select from Drug Hierarchy or Select from Event Hierarchy, and a drug or event hierarchy account is specified at the level of the data configuration, you must select a hierarchy level to tell the application which level of the hierarchy is represented by the variable.

Note:
This field has no effect if there is a hierarchy table specified for the individual variable.
If you do not want the variable to be available during creation of a query, from the Hide from Query Wizard drop-down list, select Yes.

Note:
Regardless of this setting, a variable that uses a data transformation that maps values, defines cutpoints, or converts dates is not selectable in the Query Wizard.
In the Value Case field for variables whose Oracle data type is CHAR or VARCHAR2, indicate the capitalization style that is used in the source data. The application automatically converts terms that users enter in the Drug and Event Term selection fields on the Select Criteria page to the capitalization style before a match is attempted, as follows:
- Exact Match—Entered values are not converted, and must match the case of the source data exactly.
- All Upper—Values are in all upper case in the source data. Entered values are converted to all upper case.
- All Lower—Values are in all lower case in the source data. Entered values are converted to all lower case.
- Upper First Letter Each Word—The first letter of each word is capitalized in the source data. Entered values are converted so that the first letter of each word is in upper case and the rest are in lower case. For example, the entered value, Renal failure, becomes Renal Failure.
- Upper First Letter Each Term—Values are in mixed case in the source data. Entered values are converted so that the first letter of each term is in upper case and the rest of the term is lower case. For example, the entered value, Drug Hypersensitivity, becomes Drug hypersensitivity.
In the Unique Values Table field, identify an Oracle table or view that contains the list of all possible values for the variable. You can specify an existing unique values table (from the Oracle account in which the configuration resides) by typing in its name.
1. To select from a list of tables, click Select Table.
2. From the Select Table dialog box, select the table and click Save.
  
  If you leave this field blank, the application creates a unique values table during configuration validation.
  
  Note:
  In a timestamped database, no start and end dates are associated with the unique values table, so all unique values that ever existed are in the tables.
To specify a hierarchy table for the variable, in the Hierarchy Table field, type the table name, or click Select Table to choose from all tables in the Oracle account to which you are connected. To supply a table name, click Create New Table.

The recommended way to set up event or drug hierarchies is to associate an Oracle account of hierarchy information with the configuration (on the Edit Configuration Details page). The ability to associate an individual variable with a hierarchy table still exists to provide backward compatibility with data configurations created in previous releases. For more information, see Event and drug hierarchies.
In the Data Transformation field, specify a data transformation to transform source data into other values available for selection when data mining runs are created:
- If you do not want to transform source data for the variable, click None (the default).
- To map text values, click Text mapping. This option applies only to variables that are based on a CHAR or VARCHAR value in the source database.
- To define variable cutpoints, click User-defined cutpoints. This option applies only to variables that are based on a NUMBER or DATE value in the source database.
- To exclude synthetic values, such as the names of Standardized MedDRA Queries (SMQs), click Synthetic values. This option applies only to variables that are based on a CHAR or VARCHAR value in the source database.
- To convert date values, select Date function and enter the function in the field. To select a function from a list, click Select Function. This option applies only to variables that are based on a DATE value in the source database.
Click Save.

If a warning message appears, as when there are leading or trailing white spaces in the source data for the variable, you can fix the problem or click Ignore Warnings, and then Continue.
Next, validate the data configuration. For more information, see Validate data configurations.

Parent topic: Modify a data configuration