Null or MISSING Values: MISSING

In this section:

Using a Missing Value

How to:

Specify a Missing Value

Reference:

Usage Notes for MISSING

If a segment instance exists but no data has been entered into one of its fields, that field has no value. Some types of data sources represent this absence of data as a blank space ( ) or zero (0), but others explicitly indicate an absence of data with a null indicator or as a special null value. Null values (sometimes known as missing data) are significant in reporting applications, especially those that perform aggregating functions, such as averaging.

If your type of data source supports missing data, as do FOCUS data sources and most relational data sources, then you can use the optional MISSING attribute to enable null values to be entered into and read from a field. MISSING plays a role when you:

Create new segment instances. If no value is supplied for a field for which MISSING has been turned ON in the Master File or in a DEFINE or COMPUTE definition, then the field is assigned a missing value.
Generate reports. If a field with a null value is retrieved, the field value is not used in aggregating calculations, such as averaging and summing. If the report calls for the field value to display, a special character appears to indicate a missing value. The default character is a period (.), but you can change it to any character string you wish using the SET NODATA command or the SET HNODATA command for HOLD files, as described in the Developing Applications manual.

Syntax: How to Specify a Missing Value

MISSING = {ON|OFF}

where:

ON: Distinguishes a missing value from an intentionally entered blank or zero when creating new segment instances and reporting.
OFF: Does not distinguish between missing values and blank or zero values when creating new segment instances and reporting. OFF is the default value.

Reference: Usage Notes for MISSING

Note the following rules when using MISSING:

Alias. MISSING does not have an alias.
Value. It is recommended that you set the MISSING attribute to match the field predefined null characteristic (whether the characteristic is explicitly set when the data source is created, or set by default). For example, if a relational table column has been created with the ability to accept null data, describe the field with the MISSING attribute set to ON so that its null values are correctly interpreted.
FOCUS data sources also support MISSING=ON, which assigns sets an internal flag for missing values.
Changes. You can change the MISSING attribute at any time. Note that changing MISSING does not affect the actual stored data values that were entered using the old setting. However, it does affect how that data is interpreted. If null data is entered when MISSING is turned ON, and then MISSING is switched to OFF, the data originally entered as null is interpreted as blanks (for alphanumeric fields) or zeroes (for numeric fields).

Using a Missing Value

Consider the field values shown in the following four records:

If you average these values without declaring the field with the MISSING attribute, a value of zero is automatically be supplied for the two blank records. Thus, the average of these four records is (0+0+1+3)/4, or 1. If you turn MISSING to ON, the two blank records are not used in the calculation, so the average is (1+3)/2, or 2.

Missing values in a unique segment are also automatically supplied with a zero, a blank, or a missing value depending on the MISSING attribute. What distinguishes missing values in unique segments from other values is that they are not stored. You do have to supply a MISSING attribute for fields in unique segments on which you want to perform counts or averages.

The Creating Reports manual contains a more thorough discussion of using null values (sometimes called missing data) in reports. It includes alternative ways of distinguishing these values in reports, such as using the WHERE phrase with MISSING selection operators, and creating virtual fields using the DEFINE FILE command with the SOME or ALL phrase.