DEFINE FIELD command: Difference between revisions

From m204wiki
Jump to navigation Jump to search
Line 364: Line 364:
===Usage notes===
===Usage notes===
<ul>
<ul>
<li>OI <var>CHUNK</var> fields must be <var>ORDERED NUMERIC INVISIBLE</var>. </li>
<li>OI <var>CHUNK</var> fields must be <var>ORDERED NUMERIC [[Field_design#VISIBLE_and_INVISIBLE_attributes|INVISIBLE]]</var>. </li>


<li>OI <var>CHUNK</var> fields can only be defined for <var>ORDERED NUMERIC VISIBLE</var> fields. </li>
<li>OI <var>CHUNK</var> fields can only be defined for <var>ORDERED NUMERIC VISIBLE</var> fields. </li>

Revision as of 12:16, 12 February 2019

Summary

Status
Generally Available in Model 204 version 7.5:
  • Fields as members of field groups
  • Field attributes specific to FILEORG X'100' files
  • Ordered index CHUNK field attribute
Privileges
Any user
Function
Defines the names and attributes of the fields in a Model 204 file

Syntax

For a field

[IN fileName] DEFINE [FIELD] name [(attribute [attribute]...)]

or

[IN fileName] DEFINE [FIELD] name WITH [attribute [attribute]...]


For a field within a field group

DEFINE [FIELD] name [(attribute [attribute]...)] WITH FIELDGROUP [fieldgroupname | *]

or

DEFINE [FIELD] name WITH FIELDGROUP [fieldgroupname | *] [AND attribute [attribute]...]

Where:

IN fileName Specifies the file within which the field is defined. This may not be a Model 204 group; it must be a single file. If this clause is not present, the default file, as given by the CURFILE parameter, is the file in which the field is defined.

The file in which the field is defined must be OPEN, with file manager privileges.

name or fieldgroupname (1 to 255 characters) must begin with a letter and can contain any alphanumeric characters (including space) except:

??

?$

?&

@ (or an alternative delete character)

# (or an alternative flush character)

 ;

name or fieldgroupname must not start with a reserved word or character. An embedded reserved word or character must be part of a string preceded and followed by a single quotation mark.

attribute One of the field characteristics listed in the table below. Note that each attribute has a default. Therefore, when selected, the attribute need be specified only if it differs from the default. If no attributes are specified, the field is defined with all the defaults. See Field design for more information on field attributes.


Field attributes
Attribute Abbreviation Default
AT-MOST-ONE

REPEATABLE

ONE

REPT

REPT (except in 1NF files)
BINARY BIN Not a binary field
BINARY-LARGE-OBJECT BLOB Not binary large object
CHARACTER-LARGE-OBJECT CLOB Not character large object
CODED

NON-CODED

COD

NCOD

NON-CODED
DATETIME DT Data format YYYYMMDDHHMMSSXXXXXX
(the nearest millionth of a second)
DEFERRABLE

NON-DEFERRABLE

DEF

NDEF

DEFERRABLE
FEW-VALUED

MANY-VALUED

FV

MV

MANY-VALUED
FLOAT FL Not a floating point field
FRV (for-each-value)

NON-FRV

--

NFRV

NON-FRV

KEY

NON-KEY

--

NKEY

NON-KEY

LENGTH m LEN No preallocated length
LEVEL k LVL No level
LRESERVE LRES 15
NRESERVE NRES 15
NUMERIC RANGE

NON-RANGE

RANGE, NR

NNR

NON-RANGE
OCCURS n OCC No preallocated occurrences
ORDERED

NON-ORDERED

ORD

NON-ORD

Non-ordered field
ORDERED CHARACTER ORD CHAR --
ORDERED NUMERIC ORD NUM --
SPLITPCT SPLT 50
IMMED IMM 1
PAD character -- PAD X'00'
STRING

BINARY

STR

BIN

STRING
UPDATE IN PLACE

UPDATE AT END

UP

UE

UPDATE IN PLACE
VISIBLE

INVISIBLE

VIS

INV

VISIBLE
UNIQUE

NON-UNIQUE

UNIQ

NUNIQ

NON-UNIQUE
FIELDGROUP fieldgroupname The containing field group. When fieldgroupname is specified as an asterisk (*), it means that the field can be included in any field group as well as existing outside of any field group.

Conflicting attributes, listed in the table below, cannot be specified together.

Conflicting field attributes
This attribute... Cannot be specified with the attribute(s)...
BINARY, BLOB, CLOB, or CODED LENGTH
CHUNK CONCATENATION-OF, COUNT-OCCURRENCES-OF, CREATE-TIME, CREATE-TIMEUTC, CREATE-USER, UPDATE-TIME, UPDATE-TIMEUTC, UPDATE-USER
CONCATENATION-OF BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, DATETIME, DEFAULT-VALUE, OCCURS > 1, REPEATABLE, STORE-DEFAULT, STORE-NULL, UPDATE AT END
COUNT-OCCURRENCES-OF BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, DATETIME, DBCS, DEFAULT-VALUE, INVISIBLE, OCCURS > 1, REPEATABLE, STORE-DEFAULT, STORE-NULL, UNIQUE, UPDATE AT END
CREATE-TIME BINARY, BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, DBCS, FLOAT, INVISIBLE, NUMERIC RANGE, OCCURS > 1, REPEATABLE, STORE-DEFAULT, STORE-NULL, STRING, UNIQUE, UPDATE AT END
CREATE-TIMEUTC BINARY, BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, FLOAT, INVISIBLE, NOT UNIQUE, NUMERIC RANGE, OCCURS > 1, REPEATABLE, STORE-DEFAULT, STORE-NULL, STRING, UPDATE AT END
CREATE-USER BINARY, BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, DATETIME, FLOAT, INVISIBLE, NUMERIC RANGE, OCCURS > 1, STORE-DEFAULT, STORE-NULL, UNIQUE, UPDATE AT END
DATETIME BINARY, BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, CODED, CONCATENATION-OF, CREATE-USER, KEY, OCCURS, STRING, UPDATE AT END, UPDATE-USER
DATETIME-GE, DATETIME-GT, DATETIME-LE, DATETIME-LT CODED, CONCATENATION-OF, CREATE-USER, KEY, OCCURS, UPDATE AT END, UPDATE-USER
DEFAULT-VALUE REPEATABLE, NO-DEFAULT-VALUE, STORE-NULL LITERAL, STORE-NULL NONE
ESCAPE SEPARATOR NONE
EXACTLY-ONE CONCATENATION-OF, INVISIBLE, STORE-DEFAULT LITERAL, OCCURS, STORE-NULL LITERAL, UNIQUE
FEW-VALUED MANY-VALUED
FIELDGROUP FRV, INVISIBLE, NUMERIC RANGE, OCCURS
FLOAT BINARY, INVISIBLE, NUMERIC RANGE, and STRING
FLOAT-GE, FLOAT-GT, FLOAT-LE, FLOAT-LT, INTEGER-GE, INTEGER-GT, INTEGER-LE, INTEGER-LT Other range attributes
FRV NON-KEY
IMMED NON-ORDERED
INVISIBLE NON-KEY, NON-RANGE, and NON-ORDERED;

UPDATE IN PLACE or UPDATE AT END

LENGTH-EQ Any attribute that is automatically maintained by the system; BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, LENGTH-GE, LENGTH-LE
LENGTH-GE, LENGTH-LE Any attribute that is automatically maintained by the system; BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, LENGTH-EQ
LIKE Any attribute that is automatically maintained by the system; BINARY LARGE OBJECT, CHARACTER LARGE OBJECT
LRESERVE NON-ORDERED
MANY-VALUED FEW-VALUED
MINLOBE OCCURS
NO-DOMAIN-CONSTRAINTS Any constraint attribute
NO-DEFAULT-VALUE DEFAULT-VALUE, STORE-DEFAULT
NON-CODED and NON-FRV MANY-VALUED or FEW-VALUED
NON-KEY, NON-ORDERED, and NON-RANGE DEFERRABLE or NON-DEFERRABLE
NRESERVE NON-ORDERED
NUMERIC RANGE OCCURS n (if n > 1)
OCCURS INVISIBLE
ORDERED FRV
ORDERED NUMERIC NUMERIC RANGE
PAD character LENGTH
SPLITPCT NON-ORDERED
STORE-DEFAULT CONCATENATION-OF, COUNT-OCCURRENCES-OF, CREATE-TIME, CREATE-USER, UPDATE-TIME, UPDATE-USER
STORE-NULL BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, CONCATENATION-OF, COUNT-OCCURRENCES-OF, CREATE-TIME, CREATE-USER, UPDATE-TIME, UPDATE-USER
UNIQUE DEFERRABLE
UPDATE-TIME BINARY, BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, DBCS, FLOAT, INVISIBLE, NUMERIC RANGE, OCCURS > 1, STORE-DEFAULT, STORE-NULL, REPEATABLE, UNIQUE, UPDATE AT END, UPDATE-TIMEUTC
UPDATE-TIMEUTC BINARY, BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, DBCS, FLOAT, INVISIBLE, NUMERIC RANGE, OCCURS > 1, STORE-DEFAULT, STORE-NULL, REPEATABLE, UNIQUE, UPDATE AT END
UPDATE-USER BINARY, BINARY LARGE OBJECT, CHARACTER LARGE OBJECT, DATETIME, FLOAT, INVISIBLE, NUMERIC RANGE, OCCURS > 1, STORE-DEFAULT, STORE-NULL, UNIQUE, UPDATE AT END

Paired attributes, listed in the table below, must be specified together.

Paired field attributes
This attribute... Must be specified with the attribute(s)...
CHUNK ORDERED NUMERIC and INVISIBLE
CONCATENATION-OF AT-MOST-ONE, EXACTLY-ONE, or OCCURS 1
COUNT-OCCURRENCES-OF AT-MOST-ONE, EXACTLY-ONE, or OCCURS 1
CREATE-TIME AT-MOST-ONE
CREATE-TIMEUTC AT-MOST-ONE
DEFAULT-VALUE AT-MOST-ONE or EXACTLY-ONE
ESCAPE CONCATENATION-OF
FEW-VALUED FRV or CODED
FLOAT LENGTH
MANY-VALUED FRV or CODED
MINLOBE BINARY LARGE OBJECT or CHARACTER LARGE OBJECT
OCCURS CODED, BINARY, or LENGTH
PAD LENGTH
SEPARATOR CONCATENATION-OF
STORE-DEFAULT DEFAULT-VALUE, AT-MOST-ONE, or EXACTLY-ONE
UNIQUE ORDERED
UPDATE-TIME, UPDATE-TIMEUTC AT-MOST-ONE

Syntax notes

Attributes must be separated by commas or by one or more blanks.

You can specify DEFINE FIELD only in file context.

For general syntax and usage notes that apply to all forms of the DEFINE command, see DEFINE command.

Example

DEFINE FIELD CUSTID WITH KEY FRV RANGE DEFINE FIELD AGE WITH KEY RANGE OCCURS 1 LENGTH 3 DEFINE FIELD EMPLOYEE

Usage notes

The DEFINE FIELD command describes the names and attributes of the fields in a Model 204 file. The file must be initialized, using the INITIALIZE command, before fields can be defined.

Once fields have been defined, a user with the requisite privileges can perform the following commands on them:

Record security and sort and hash key fields

Sort key fields for sorted files, hash key fields, and record security fields can be defined only with the INITIALIZE command, not with the DEFINE FIELD command.

Floating point fields

You can define three types of floating point fields through the FLOAT and LENGTH attributes. The table below shows the characteristics of the three types.

Types of floating point fields
Type Number of bytes Maximum number of significant digits
Short precision 4 6
Long precision 8 15
Extended precision 16 31

You supply a LENGTH attribute of 4, 8, or 16. The following example defines a long-precision floating-point field:

DEFINE FIELD SALARY WITH FLOAT LEN 8

Consider making a field a floating-point field if its value is always or often numerical. Very large numbers or numbers with many significant digits use much less space in floating-point form than they do in character form. Floating-point fields are valid only for files with the format of Release 8.0 or later or that have been converted by the TRANSFORM FLOAT command. Floating-point numbers in ORDERED NUMERIC trees are discussed as follows.

Defining a field group field

A field defined within a field group supports the same default attributes as fields defined outside or independent of a field group, except for REPEATABLE. A field defined within a field group defaults to EXACTLY-ONE. You can override the default with REPEATABLE. The attribute for frequency of occurrence of the field being is one of these:

  • EXACTLY-ONE, the default for a field defined within a DEFINE FIELDGROUP block
  • REPEATABLE, the default for a field defined outside of a field group
  • AT-MOST-ONE

ORDERED attribute

The ORDERED attribute is used to define fields for the Ordered Index feature. The ORDERED attribute can be followed by a tree type to specify the kind of ordering for the index. The tree type can be either CHARACTER (CHAR) or NUMERIC (NUM). If a tree type is not provided, a default tree type is used.

Syntax

The format for the DEFINE FIELD command with solely the ORDERED option and its related attributes is as follows:

DEFINE [FIELD] name WITH [ORDERED [CHARACTER | NUMERIC] [LRESERVE n] [NRESERVE n] [SPLITPCT n] [IMMED n]]

Where:

LRESERVE The percentage of space to leave free on the leaf page during deferred updates or a REORGANIZE command. The default is 15 percent. The value must be between 0 and 99.
NRESERVE The percentage of space to leave free on the node page during deferred updates or a REORGANIZE command. The default is 15 percent. The value must be between 0 and 99.
SPLITPCT The percentage of the node data to go to the left node when a split occurs during nondeferred SOUL or IFAM requests. The default is 50 percent. The value must be between 1 and 100.
IMMED The number of immediate pointers per segment in the Ordered Index. The default is 1. The value must be between 0 and 255.

Usage notes

The LRESERVE, NRESERVE, SPLITPCT, and IMMED parameters represent approximate values. The actual movement of data during an update might not conform exactly to the value of the attribute in question.

The default ordering for a field is NON-ORDERED. Ordered Index attributes do not apply for a NON-ORDERED field.

If a field is defined as ORDERED without specifying the type of tree to create, the default depends on whether the field is STRING, BINARY, or FLOAT. If STRING, an ORDERED CHARACTER tree is defined. If FLOAT or BINARY, an ORDERED NUMERIC tree is defined.

The UNIQUE attribute specifies that there can be only one occurrence of a unique field in the file. Multiple occurrences are not allowed.

For a detailed discussion of field attributes, see Field design.

Ordered index CHUNK attribute

The Ordered Index CHUNK field attribute is available in Model 204 version 7.5. It improves the efficiency of range finds using Ordered Index (OI) processing.

The OI CHUNK attribute enables more efficient searching on Ordered Index NUMERIC RANGE (ORDERED NUMERIC) fields by specifying "OI chunks" of data. CHUNK is used to define a related field containing data from the original field (the "target" field) rounded down to the OI chunk size.

Syntax

The format for the DEFINE FIELD command with the CHUNK option and its related attributes is as follows.

Note that the CHUNK keyword can be abbreviated CNK.

DEFINE FIELD chunkFieldName WITH INVISIBLE ORDERED NUMERIC {CHUNK|CNK} chunkSize FOR chunkTargetFieldName

Where:

chunkFieldName A standard Model 204 field name.
chunkSize A positive integer that is an integral multiple or divisor of all OI CHUNK fields previously defined for chunkTargetFieldName.

The maximum OI chunk size is 2,147,483,647.

OI chunk size also cannot be the same as the OI chunk size for any other CHUNK field defined for chunkTargetFieldName. If either OI chunk size rule is violated, the DEFINE command is rejected with an error message.

chunkTargetFieldName A previously defined ORDERED NUMERIC VISIBLE field (it must be defined before any CHUNK fields that reference it). The DEFINE command is rejected with an error message if the CHUNK target is not ORDERED NUMERIC VISIBLE.

Usage notes

  • OI CHUNK fields must be ORDERED NUMERIC INVISIBLE.
  • OI CHUNK fields can only be defined for ORDERED NUMERIC VISIBLE fields.
  • Any given OI CHUNK target field can have at most 20 CHUNK fields defined for it.
  • In version 7.5, OI CHUNK target fields cannot be field group members.
    As of version 7.6, OI CHUNK target fields can be members of field groups.
  • OI CHUNK fields are an Automatic field type. Therefore:
    • The CHUNK type is mutually exclusive with all other automatic field types (CONCATENATION-OF, COUNT-OCCURRENCES-OF, CREATE-TIME, CREATE-TIMEUTC, CREATE-USER, UPDATE-TIME, UPDATE-TIMEUTC, UPDATE-USER).
    • Field values for OI CHUNK fields are set internally as instances of the CHUNK target field are added, deleted, and changed. The value of a CHUNK field is updated at the start of a transaction by Model 204, and it cannot be set explicitly by a program.
  • Once you have defined an OI CHUNK field, you cannot redefine it, rename it, or delete it.

File restrictions when using OI CHUNK fields

  • CHUNK fields can only be defined in a FILEORG X'100' file.
  • Once a file has CHUNK fields defined, it cannot be opened with a Model 204 version prior to 7.5.
  • A CHUNK field cannot be defined in a non-empty file.

OI CHUNK target field value restrictions

An Ordered Index CHUNK target field can only store values between -999999999999999 and 999999999999999 (15 digit signed integer). If an attempt is made to store a value outside of that range, the request is cancelled with the following error:

CANCELLING REQUEST: M204.1402: Invalid value for CHUNK target: target-field-name = bad-value; values must be -999999999999999->999999999999999

DATETIME fields used as OI CHUNK targets

A DATETIME field that also has the ORD NUM (ORDERED NUMERIC) attribute can be used as a CHUNK target. Remember that independent of OI chunks, only the 15 leftmost digits of a DATETIME field's 20 digits are stored in the ORD NUM index. When computing the values of OI chunks for a DATETIME target, base the computation on 15 digits, not 20. For example:

DEFINE TIMESTAMP WITH DT ORD NUM DEFINE TIMESTAMP_YYYYMM WITH ORD NUM INV CHUNK 1000000000 FOR TIMESTAMP DEFINE TIMESTAMP_YYYYMMDD WITH ORD NUM INV CHUNK 10000000 FOR TIMESTAMP DEFINE TIMESTAMP_YYYYMMDDHH WITH ORD NUM INV CHUNK 100000 FOR TIMESTAMP DEFINE TIMESTAMP_YYYYMMDDHHMM WITH ORD NUM INV CHUNK 1000 FOR TIMESTAMP

This rule also applies to the automatic date field types (UPTM, UPTMU, CRTM, CRTMU).

Date/time stamp field

Specifying the date/time stamp field name

The system parameter DTSFN specifies the date/time stamp field name. This field name is the same for all files participating in DTS processing in a run. Though this is a system-level parameter, the feature is a file-level feature.

The date/time stamp field name you specify is not edited for validity.

The DTSFN value can be specified in CUST source code so you do not have to change all your jobs to specify the parameter. By default, Model 204 does not provide a date/time stamp field name.

Specifying a date/time stamp field definition

A date/time stamp field is defined as is any other Model 204 field but has the following field attribute restrictions:

  • A date/time field cannot be INVISIBLE or UNIQUE.
  • If the date/time stamp field is preallocated, OCCURS cannot be greater than 1.

Disregarding these restrictions causes the following error:

M204.2728: DATE TIME STAMP FIELD MAY NOT BE INVISIBLE, UNIQUE, NOR HAVE MULTIPLE OCCURRENCES

Rocket Software recommends using the ORDERED CHARACTER attribute. By default Model 204 generates a character string date/time stamp field value.

Defining the date/time stamp field value format

Your custom code formats the data entered into the field. The format for the default date/time stamp string is:

CCYYMMDDHHMMSSMMM

Note: The DTSFN field can have differing definitions in different Model 204 databases.