Describing a Single Segment

FOCUS data sources use the SEGTYPE attribute to describe segment key fields and sort order, as well as the relationship of the segment to its parent.

The SEGTYPE attribute is also used with SUFFIX=FIX data sources to indicate a logical key sequence for that data source. SEGTYPE is discribed in Describing a Group of Fields.

Use the SEGTYPE attribute to describe which fields in a segment are key fields. The values of these fields determine how the segment instances are sequenced. The keys must be the first fields in a segment. You can specify up to 255 keys in a segment that is sorted from low to high (SEGTYPE = Sn), and up to 99 keys in a segment sorted from high to low (SEGTYPE = SHn). To maximize efficiency, it is recommended that you specify only as many keys as you need to make each record unique. You can also choose not to have any keys (SEGTYPE = S0 and SEGTYPE = blank).

Note: Text fields cannot be used as key fields.

For segments that have key fields, use the SEGTYPE attribute to describe the segment sort order. You can sort a segment instances in two ways:

• Low to high. By specifying a SEGTYPE of Sn (where n is the number of keys), the instances are sorted using the concatenated values of the first n fields, beginning with the lowest value and continuing to the highest.
• High to low. By specifying a SEGTYPE of SHn (where n is the number of keys), the instances are sorted using the concatenated values of the first n fields, beginning with the highest value and continuing to the lowest.

  Segments whose key is a date field often use a high-to-low sort order, since it ensures that the segment instances with the most recent dates are the first ones encountered in a segment chain.

Suppose the following fields in a segment represent a department code and the employee last name:

If you set SEGTYPE to S1, the department code becomes the key. Note that two records have duplicate key values in order to illustrate a point about S2 segments later in this example. Duplicate key values are not recommended for S1 and SH1 segments. The segment instances are sorted as follows:

If you change the field order to put the last name field before the department code and leave SEGTYPE as S1, the last name becomes the key. The segment instances are sorted as follows:

Alternately, if you leave the department code as the first field, but set SEGTYPE to S2, the segments are sorted first by the department code and then by last name, as follows:

The SEGTYPE attribute describes the relationship of a segment to its parent segment:

• Physical one-to-one relationships are usually specified by setting SEGTYPE to U. If a segment is described in a Master File-defined join as descending from the cross-referenced segment, then SEGTYPE is set to KLU in the join description.
• Physical one-to-many relationships are specified by setting SEGTYPE to any valid value beginning with S (such as S0, SHn, and Sn) to blank, or, if a segment is described in a Master File-defined join as descending from the cross-referenced segment, to KL.
• One-to-one joins defined in a Master File are specified by setting SEGTYPE to KU or DKU, as described in Defining a Join in a Master File.
• One-to-many joins defined in a Master File are specified by setting SEGTYPE to KM or DKM, as described in Defining a Join in a Master File.

By default, all of the segments in a FOCUS data source are stored in one physical file. For example, all of the EMPLOYEE data source segments are stored in the data source named EMPLOYEE.

Use the LOCATION attribute to specify that one or more segments be stored in a physical file separate from the main data source file. The LOCATION file is also known as a horizontal partition. You can use a total of 64 LOCATION files per Master File (one LOCATION attribute per segment, except for the root). This is helpful if you want to create a data source larger than the FOCUS limit for a single data source file, or if you want to store parts of the data source in separate locations for security or other reasons.

There are at least two cases in which to use the LOCATION attribute:

• Each physical file is individually subject to a maximum file size. You can use the LOCATION attribute to increase the size of your data source by splitting it into several physical files, each one subject to the maximum size limit. (See Defining a Join in a Master File to learn if it is more efficient to structure your data as several joined data sources.)
• You can also store your data in separate physical files to take advantage of the fact that only the segments needed for a report must be present. Unreferenced segments stored in separate data sources can be kept on separate storage media to save space or implement separate security mechanisms. In some situations, separating the segments into different data sources allows you to use different disk drives.

Divided data sources require more careful file maintenance. Be especially careful about procedures that are done separately to separate data sources, such as backups. For example, if you do backups on Tuesday and Thursday for two related data sources, and you restore the FOCUS structure using the Tuesday backup for one half and the Thursday backup for the other, there is no way of detecting this discrepancy.

Text fields, by default, are stored in one physical file with non-text fields. However, as with segments, a text field can be located in its own physical file, or any combination of text fields can share one or several physical files. Specify that you want a text field stored in a separate file by using the LOCATION attribute in the field definition.

For example, the text for DESCRIPTION is stored in a separate physical file named CRSEDESC:

FIELD = DESCRIPTION, ALIAS = CDESC, USAGE = TX50, LOCATION = CRSEDESC ,$

If you have more than one text field, each field can be stored in its own file, or several text fields can be stored in one file.

In the following example, the text fields DESCRIPTION and TOPICS are stored in the LOCATION file CRSEDESC. The text field PREREQUISITE is stored in another file, PREREQS.

FIELD = DESCRIPTION , ALIAS = CDESC, USAGE = TX50, LOCATION = CRSEDESC,$
FIELD = PREREQUISITE, ALIAS = PREEQ, USAGE = TX50, LOCATION = PREREQS ,$
FIELD = TOPICS,       ALIAS =      , USAGE = TX50, LOCATION = CRSEDESC,$

As with segments, you might want to use the LOCATION attribute on a text field if it is very long. However, unlike LOCATION segments, LOCATION files for text fields must be present during a request, whether or not the text field is referenced.

The LOCATION attribute can be used independently for segments and for text fields. You can use it for a text field without using it for a segment. You can also use the LOCATION attribute for both the segment and the text field in the same Master File.

Note: Field names for text fields in a FOCUS Master File are limited to 12 characters. Field names for text fields in an XFOCUS Master File are not subject to this 12 character limitation. However, for both types of data sources, alias names for these fields can be up to 66 characters.

The maximum number of segments in a Master File is 64. There is a limit on the number of different location segments and text LOCATION files you can specify. This limit is based on the number of entries allowed in the File Directory Table (FDT) for FOCUS and XFOCUS data sources. The FDT contains the names of the segments in the data source, the names of indexed fields, and the names of LOCATION files for text fields.

In addition to specifying a DATASET attribute at the file level in a FOCUS Master File, you can specify the attribute on the segment level to specify the physical file name for a LOCATION segment, or a cross-referenced segment with field redefinitions.

For information on specifying the DATASET attribute at the file level, see Identifying a Data Source.

The segment with the DATASET attribute must be either a LOCATION segment or a cross-referenced segment. For cross-referenced segments:

• If field declarations are specified for the cross-referenced fields, the DATASET attribute is the only method for specifying a physical file, because the cross-referenced Master File is not read and therefore is not able to pick up its DATASET attribute if one is specified.
• If field declarations are not specified for the cross-referenced fields, it is better to place the DATASET attribute at the file level in the cross-referenced Master File. In this case, specifying different DATASET values at the segment level in the host Master File and the file level of the cross-referenced Master File causes a conflict, resulting in a (FOC1998) message.

If DATASET is used in a Master File whose data source is managed by the FOCUS Database Server, the DATASET attribute is ignored on the server side because the FOCUS Database Server does not read Master Files for servicing table requests.

The DATASET attribute in the Master File has the lowest priority:

• A user explicit allocation overrides DATASET attributes.
• The USE command for FOCUS data sources overrides DATASET attributes and explicit allocations.

Note: If a DATASET allocation is in effect, you must issue a CHECK FILE command in order to override it by an explicit allocation command. The CHECK FILE command deallocates the allocation created by DATASET.

Each segment of a FOCUS data source can have a timestamp field that records the date and time of the last change to the segment. This field can have any name, but its USAGE format must be AUTODATE. The field is populated each time its segment instance is updated. The timestamp is stored as format HYYMDS, and can be manipulated for reporting purposes using any of the date-time functions.

In each segment of a FOCUS data source, you can define a field with USAGE = AUTODATE. The AUTODATE field cannot be part of a key field for the segment. Therefore, if the SEGTYPE is S2, the AUTODATE field cannot be the first or second field defined in the segment.

The AUTODATE format specification is supported only for a real field in the Master File, not in a DEFINE or COMPUTE command or a DEFINE in the Master File. However, you can use a DEFINE or COMPUTE command to manipulate or reformat the value stored in the AUTODATE field.

After adding an AUTODATE field to a segment, you must REBUILD the data source. REBUILD does not timestamp the field. It does not have a value until a segment instance is inserted or updated.

If a user-written procedure updates the AUTODATE field, the user-specified value is overwritten when the segment instance is written to the data source. No message is generated to inform the user that the value was overwritten.

The AUTODATE field can be indexed. However, it is recommended that you make sure the index is necessary, because of the overhead needed to keep the index up to date each time a segment instance changes.

If you create a HOLD file that contains the AUTODATE field, it is propagated to the HOLD file as a date-time field with the format HYYMDS.