Additional information available:

ALIGNEDARRAYBASEBITBLANK_WHEN_ZEROBYTE
COLUMN_MAJORCOMPUTED_BY_DATATRIEVECONDITION_NAME
COPY_FIELDDATATYPEDATEDEFAULT_VALUEDESCRIPTION
D_FLOATINGD_FLOATING_COMPLEXEDIT_CODE
EDIT_STRINGEDIT_WORDELEMENTARY_FIELDFIELD_DESCRIPTION_CLAUSE
F_FLOATINGF_FLOATING_COMPLEXG_FLOATING
G_FLOATING_COMPLEXHELPH_FLOATING COMPLEX
INITIAL_VALUEJUSTIFIED_RIGHTLEFT_OVERPUNCHED_NUMERIC
LEFT_SEPARATE_NUMERICLONGWORDNAMEOCCURS_DEPENDING
OCTAWORDPACKED_DECIMALPICTUREPOINTERQUADWORD
QUERY_HEADERQUERY_NAMERIGHT_OVERPUNCHED_NUMERIC
RIGHT_SEPARATE_NUMERICROW_MAJORSCALESIZE
STRUCTURE_FIELDTEXTUNSIGNED_NUMERICUNSPECIFIED
VALID_FOR_DATATRIEVE_IFVARYING_STRINGVIRTUAL_FIELD
WORDZONED_NUMERIC

bit field sizeboundary optionsdatatype optionsfield attribute statement
field description statementfield nameFixed pointFloating pointlower bound
no digitsno fractspath nameradixRecord definitionscale factor
text field sizetext stringunspecified field sizeupper boundVARIANTS field
VARIANTS OF fieldvs field size

 An unsigned integer indicating the number of bits in the field.

 Location where field is  aligned  relative  to  start  of  record  in
 ALIGNED clause.

 The DATATYPE clause allows  you  several  options  for  describing  a
 field's data type:

Additional information available:

DATEVIRTUAL_FIELDBITUNSPECIFIEDTEXT
VARYING STRINGPOINTER

Floating pointFixed pointDecimal string

 Specifies that the field is a 64-bit VAX standard absolute date  data
 type.

 Specifies that the field is a VAX DATATRIEVE virtual field.  No space
 is  allocated  for  virtual  fields  in  a  record.   The COMPUTED BY
 DATATRIEVE clause determines the value of  a  virtual  field  at  run
 time.  A STRUCTURE field cannot contain the VIRTUAL FIELD datatype.

 Specifies that the field is a bit string.   Indicate  the  number  of
 bits in the field with an unsigned integer.

 Declares that the field  is  a  sequence  of  8-bit  unsigned  bytes.
 Indicate the number of bytes in the field with an unsigned integer.

 Specifies that  the  field  is  a  sequence  of  8-bit  ASCII  bytes.
 Indicate  the  number  of  characters  in  the field with an unsigned
 integer.  CDDL accepts CHARACTER as a synonym for TEXT.

 Specifies that  the  field  is  a  PL/I  or  PASCAL  varying  string.
 Indicate  the  number  of  characters  in  the field with an unsigned
 integer.  CDDL accepts VARYING TEXT as a synonym for VARYING STRING.

 Specifies that the field contains the address  of  another  field  or
 record  definition.  In PL/I, for example, POINTER fields are used to
 access based variables and buffers allocated by the system.  Although
 PL/I  does  not  associate  POINTER  fields  with  a specified record
 structure, other languages do; the optional [TO path-name]  lets  you
 connect a POINTER to a structure.

 Floating point data types represent approximations to quantities in a
 scientific  notation  consisting  of a signed integer and a fraction.
 The four floating point data types are  the  32-bit  F_FLOATING,  the
 64-bit D_FLOATING, the 64-bit G_FLOATING, and the 128-bit H_FLOATING.
 In addition, you can  use  F_FLOATING  COMPLEX,  D_FLOATING  COMPLEX,
 G_FLOATING  COMPLEX,  and H_FLOATING COMPLEX to specify ordered pairs
 of floating point data types  representing  the  real  and  imaginary
 components of complex numbers.

Additional information available:

SCALEBASED_FLOATINGD_FLOATING_COMPLEX
F_FLOATINGF_FLOATING_COMPLEXG_FLOATING
G_FLOATING_COMPLEXH_FLOATINGH_FLOATING COMPLEX

 You can specify a SCALE as an implied exponent.  The  signed  integer
 must  be in the range -128 to 127.  The SCALE specification indicates
 the number of places to shift the decimal point  when  the  field  is
 evaluated.   A  negative  integer  indicates a shift of the specified
 number of places to the left, and  a  positive  integer  indicates  a
 shift of the specified number of places to the right.

 You can also specify the radix, or BASE, with  an  unsigned  integer.
 The  BASE  indicates  the  number system to be used when the field is
 evaluated.  The default BASE is 10.

 Specifies that the field is  a  64-bit  floating  point  number  with
 precision to approximately 16 decimal digits.

 Specifies that the field consists  of  two  64-bit  floating  complex
 numbers,  one for the real component and one for the imaginary.  CDDL
 accepts D_FLOATING_COMPLEX as a synonym for D_FLOATING COMPLEX.

 Specifies that the field is  a  32-bit  floating  point  number  with
 precision to approximately seven decimal digits.

 Specifies that the field consists  of  two  32-bit  floating  complex
 numbers,  one for the real component and one for the imaginary.  CDDL
 accepts FLOATING_COMPLEX, FLOATING COMPLEX, and F_FLOATING_COMPLEX as
 synonyms for F_FLOATING COMPLEX.

 Specifies that the field is an extended range 64-bit  floating  point
 number with precision to approximately 15 decimal digits.

 Specifies that the  field  consists  of  two  extended  range  64-bit
 floating  complex numbers, one for the real component and one for the
 imaginary.   CDDL  accepts  G_FLOATING_COMPLEX  as  a   synonym   for
 G_FLOATING COMPLEX.

 Specifies that the field is an extended range 128-bit floating  point
 number with precision to approximately 33 decimal digits.

 Specifies that the field  consists  of  two  extended  range  128-bit
 floating  complex numbers, one for the real component and one for the
 imaginary.   CDDL  accepts  H_FLOATING_COMPLEX  as  a   synonym   for
 H_FLOATING COMPLEX.

 Fixed point data  types  represent  scaled  quantities  in  a  binary
 format.  They can be SIGNED or UNSIGNED.

  o  You can declare the total number of  DIGITS  and  the  number  of
     those  digits  that  are FRACTIONS.  The number of digits must be
     greater than 0 and less than 32.  The number  of  fractions  must
     not  be greater than the number of digits.  The default number of
     fractions is 0.

  o  You can specify a SCALE  as  an  implied  exponent.   The  signed
     integer   must   be   in  the  range  -128  to  127.   The  SCALE
     specification indicates the number of places to shift the decimal
     point  when  the  field  is  evaluated.  A negative specification
     indicates a shift of the specified number of places to the  left,
     and  a  positive specification indicates a shift of the specified
     number of places to the right.

     When using SCALE, note the following rules:

      -  The SCALE specification and the FRACTIONS specification  both
         indicate  how  many digits in a field are to the right of the
         decimal   point.    There   are,   however,   two   important
         differences:

          -  The number of FRACTIONS you can specify is limited to the
             number  of  DIGITS  you  declare for the data type.  With
             SCALE, there is no such limitation.

          -  With SCALE, you can shift the decimal point to the  right
             and  to  the  left, but with FRACTIONS, you can shift the
             decimal point only to the left.


      -  Use a positive integer in the SCALE specification to move the
         decimal point to the right.

      -  Use a negative integer in the SCALE specification to move the
         decimal point to the left.


  o  You can also  specify  the  radix,  or  BASE,  with  an  unsigned
     integer.   The  BASE  indicates the number system to be used when
     the field is evaluated.  The default BASE is 10.

Additional information available:

BYTEWORDLONGWORDQUADWORDOCTAWORD

Example

 Specifies that the field is an 8-bit byte.  The BYTE can be SIGNED or
 UNSIGNED.   If  there  is  no  sign  specification,  UNSIGNED  is the
 default.

 Specifies that the field is a 16-bit word.  The field can  be  SIGNED
 or UNSIGNED.  If no sign is specified, UNSIGNED is the default.

 Specifies that the field is a 32-bit longword.  The LONGWORD  can  be
 SIGNED  or  UNSIGNED.   If  no  sign  is  specified,  UNSIGNED is the
 default.

 Specifies that the field is a 64-bit quadword field.  The  field  can
 be  SIGNED  or  UNSIGNED.   If  no sign is specified, UNSIGNED is the
 default.

 Specifies that the field is a 128-bit octaword field.  The field  can
 be  SIGNED  or  UNSIGNED.   If  no sign is specified, UNSIGNED is the
 default.

 The STRUCTURE field BACK_ORDER contains examples of valid  CDDL  data
 declarations.

 BACK_ORDER STRUCTURE.
     PRODUCT_NO     DATATYPE IS TEXT
                    SIZE IS 8 CHARACTERS.
     DATE_ORDERED   DATATYPE IS DATE.
     STATUS_CODE    DATATYPE IS BYTE.
     QUANTITY       DATATYPE IS LONGWORD
                    ALIGNED ON LONGWORD.
     SUPPLIER       ARRAY 1:4
                    DATATYPE IS TEXT
                    SIZE IS 30 CHARACTERS.
     UNIT_PRICE     DATATYPE IS LONGWORD SCALE -2.
 END BACK_ORDER STRUCTURE.

  o  You must declare the total number of DIGITS in  the  field.   You
     can also declare which of those digits are FRACTIONS.  The number
     of digits must be greater than 0 and less than 32.  The number of
     fractions  must  not  be  greater than the number of digits.  The
     default number of fractions is 0.

  o  You can specify a SCALE  as  an  implied  exponent.   The  signed
     integer   must   be   in  the  range  -128  to  127.   The  SCALE
     specification indicates the number of places to shift the decimal
     point  when  the  field  is  evaluated.  A negative specification
     indicates a shift of the specified number of places to the  left,
     and  positive indicates a shift of the specified number of places
     to the right.

     When using SCALE, note the following rules:

      -  The SCALE specification and the FRACTIONS specification  both
         indicate  how  many digits in a field are to the right of the
         decimal   point.    There   are,   however,   two   important
         differences:

          o  The number of FRACTIONS you can specify is limited to the
             number  of  DIGITS  you  declare for the data type.  With
             SCALE, there is no such limitation.

          o  With SCALE, you can shift the decimal point to the  right
             and  to  the  left, but with FRACTIONS, you can shift the
             decimal point only to the left.


      -  Use a positive integer in the SCALE specification to move the
         decimal point to the right.

      -  Use a negative integer in the SCALE specification to move the
         decimal point to the left.


  o  You can also  specify  the  radix,  or  BASE,  with  an  unsigned
     integer.   The  BASE  indicates the number system to be used when
     the field is evaluated.  The default BASE is 10.

Additional information available:

PACKED DECIMALUNSIGNED NUMERICZONED NUMERICLEFT SEPARATE NUMERIC
LEFT OVERPUNCHED NUMERICRIGHT SEPARATE NUMERICRIGHT OVERPUNCHED NUMERIC

 Specifies that the field is a packed  decimal  numeric  field.   CDDL
 accepts PACKED NUMERIC as a synonym for PACKED DECIMAL.

 Specifies that the field is an unsigned decimal string.  You must use
 the UNSIGNED keyword.

 Specifies that the field is a VAX  ZONED  NUMERIC  data  type.   CDDL
 accepts SIGNED NUMERIC as a synonym for ZONED NUMERIC.

 Specifies that the field is a left separate  signed  numeric  decimal
 string.  CDDL accepts [SIGNED] NUMERIC LEFT SEPARATE as a synonym for
 LEFT SEPARATE NUMERIC.

 Specifies that the field is a left overpunched signed numeric decimal
 string.   CDDL accepts [SIGNED] NUMERIC LEFT OVERPUNCHED as a synonym
 for LEFT OVERPUNCHED NUMERIC.

 Specifies that the field is a right separate signed  numeric  decimal
 string.   CDDL  accepts  [SIGNED] NUMERIC RIGHT SEPARATE as a synonym
 for RIGHT SEPARATE NUMERIC.

 Specifies that the  field  is  a  right  overpunched  signed  numeric
 decimal string.  CDDL accepts [SIGNED] NUMERIC RIGHT OVERPUNCHED as a
 synonym for RIGHT OVERPUNCHED NUMERIC.

 In each field attribute statement, you use field attribute clauses to
 define the fields in a record.  CDDL contains generic field attribute
 clauses and some facility-specific field attribute clauses  that  are
 understood by only a particular language.

 CDDL contains five types of field description  statements.   You  can
 use  an  elementary,  STRUCTURE, COPY, VARIANTS, or VARIANTS_OF field
 description statement, but you can use  only  one  field  description
 statement to define a record.

 STRUCTURE,  COPY,  VARIANT,   and   VARIANTS_OF   field   description
 statements   are   subdivided   and   defined  by  subordinate  field
 description statements.

 The field name you assign can be up to 31  characters  from  the  set
 A-Z,  0-9,  _, and $.  The first character must be a letter from A-Z,
 and the last character can  not  be  _  or  $.   You  can  use  8-bit
 alphabetic characters in field names.

 Fixed point data  types  represent  scaled  quantities  in  a  binary
 format.  They can be SIGNED or UNSIGNED.

  o  You can declare the total number of  DIGITS  and  the  number  of
     those  digits  that  are FRACTIONS.  The number of digits must be
     greater than 0 and less than 32.  The number  of  fractions  must
     not  be greater than the number of digits.  The default number of
     fractions is 0.

  o  You can specify a SCALE  as  an  implied  exponent.   The  signed
     integer   must   be   in  the  range  -128  to  127.   The  SCALE
     specification indicates the number of places to shift the decimal
     point  when  the  field  is  evaluated.  A negative specification
     indicates a shift of the specified number of places to the  left,
     and  a  positive specification indicates a shift of the specified
     number of places to the right.

     When using SCALE, note the following rules:

      -  The SCALE specification and the FRACTIONS specification  have
         a  similar function.  They both indicate how many digits in a
         field are to the right of  the  decimal  point.   There  are,
         however, two important differences:

          o  The number of FRACTIONS you can specify is limited to the
             number  of  DIGITS  you  declare for the data type.  With
             SCALE, there is no such limitation.

          o  With SCALE, you can shift the decimal point to the  right
             and  to  the  left, but with FRACTIONS, you can shift the
             decimal point only to the left.


      -  Use a positive integer in the SCALE specification to move the
         decimal point to the right.

      -  Use a negative integer in the SCALE specification to move the
         decimal point to the left.


  o  You can also  specify  the  radix,  or  BASE,  with  an  unsigned
     integer.   The  BASE  indicates the number system to be used when
     the field is evaluated.  The default BASE is 10.

Additional information available:

BYTEWORDLONGWORDQUADWORDOCTAWORD
PACKED DECIMALUNSIGNED NUMERICZONED NUMERICLEFT SEPARATE NUMERIC
LEFT OVERPUNCHED NUMERICRIGHT SEPARATE NUMERICRIGHT OVERPUNCHED NUMERIC

 Specifies that the field is an 8-bit byte.  The BYTE can be SIGNED or
 UNSIGNED.   If  there  is  no  sign  specification,  UNSIGNED  is the
 default.

 Specifies that the field is a 16-bit word.  The field can  be  SIGNED
 or UNSIGNED.  If no sign is specified, UNSIGNED is the default.

 Specifies that the field is a 32-bit longword.  The LONGWORD  can  be
 SIGNED  or  UNSIGNED.   If  no  sign  is  specified,  UNSIGNED is the
 default.

 Specifies that the field is a 64-bit quadword field.  The  field  can
 be  SIGNED  or  UNSIGNED.   If  no sign is specified, UNSIGNED is the
 default.

 Specifies that the field is a 128-bit octaword field.  The field  can
 be  SIGNED  or  UNSIGNED.   If  no sign is specified, UNSIGNED is the
 default.


 1 Decimal_string

  o  You must declare the total number of DIGITS in  the  field.   You
     can also declare which of those digits are FRACTIONS.  The number
     of digits must be greater than 0 and less than 32.  The number of
     fractions  must  not  be  greater than the number of digits.  The
     default number of fractions is 0.

  o  You can specify a SCALE  as  an  implied  exponent.   The  signed
     integer   must   be   in  the  range  -128  to  127.   The  SCALE
     specification indicates the number of places to shift the decimal
     point  when  the  field  is  evaluated.  A negative specification
     indicates a shift of the specified number of places to the  left,
     and  positive indicates a shift of the specified number of places
     to the right.

     When using SCALE, note the following rules:

      -  The SCALE specification and the FRACTIONS specification  have
         a  similar function.  They both indicate how many digits in a
         field are to the right of  the  decimal  point.   There  are,
         however, two important differences:

          o  The number of FRACTIONS you can specify is limited to the
             number  of  DIGITS  you  declare for the data type.  With
             SCALE, there is no such limitation.

          o  With SCALE, you can shift the decimal point to the  right
             and  to  the  left, but with FRACTIONS, you can shift the
             decimal point only to the left.


      -  Use a positive integer in the SCALE specification to move the
         decimal point to the right.

      -  Use a negative integer in the SCALE specification to move the
         decimal point to the left.


  o  You can also  specify  the  radix,  or  BASE,  with  an  unsigned
     integer.   The  BASE  indicates the number system to be used when
     the field is evaluated.  The default BASE is 10.

 Specifies that the field is a packed  decimal  numeric  field.   CDDL
 accepts PACKED NUMERIC as a synonym for PACKED DECIMAL.

 Specifies that the field is an unsigned decimal string.  You must use
 the UNSIGNED keyword.

 Specifies that the field is a VAX  ZONED  NUMERIC  data  type.   CDDL
 accepts SIGNED NUMERIC as a synonym for ZONED NUMERIC.

 Specifies that the field is a left separate  signed  numeric  decimal
 string.  CDDL accepts [SIGNED] NUMERIC LEFT SEPARATE as a synonym for
 LEFT SEPARATE NUMERIC.

 Specifies that the field is a left overpunched signed numeric decimal
 string.   CDDL accepts [SIGNED] NUMERIC LEFT OVERPUNCHED as a synonym
 for LEFT OVERPUNCHED NUMERIC.

 Specifies that the field is a right separate signed  numeric  decimal
 string.   CDDL  accepts  [SIGNED] NUMERIC RIGHT SEPARATE as a synonym
 for RIGHT SEPARATE NUMERIC.

 Specifies that the  field  is  a  right  overpunched  signed  numeric
 decimal string.  CDDL accepts [SIGNED] NUMERIC RIGHT OVERPUNCHED as a
 synonym for RIGHT OVERPUNCHED NUMERIC.

 Floating point data types represent approximations to quantities in a
 scientific  notation  consisting  of a signed integer and a fraction.
 The four floating point data types are  the  32-bit  F_FLOATING,  the
 64-bit D_FLOATING, the 64-bit G_FLOATING, and the 128-bit H_FLOATING.
 In addition, you can  use  F_FLOATING  COMPLEX,  D_FLOATING  COMPLEX,
 G_FLOATING  COMPLEX,  and H_FLOATING COMPLEX to specify ordered pairs
 of floating point data types  representing  the  real  and  imaginary
 components of complex numbers.

 Specifies the lower bound of a subscript in an ARRAY field  attribute
 clause.  The default value is 1.

 The total number of digits.  No_digits must be  greater  than  0  and
 less than 32.

 The number of no_digits that are fractions.  No_fracts  must  not  be
 greater than no_digits.  The default no_fracts is 0.

 The location of the record definition in the CDD directory hierarchy.
 The  full  dictionary  path  name  of  a directory, subdictionary, or
 object consists of the concatenation of the given names  of  all  its
 ancestors,  beginning  with  CDD$TOP  and ending with its given name.
 Separate each given name from the next with a period.

 An example of a path_name is CDD$TOP.CORPORATE.ADDRESS_RECORD.

 An unsigned integer.  Indicates the number system to be used when the
 field is evaluated.  The default radix is 10.

 The DEFINE RECORD statement begins  each  record  definition  in  the
 source file.

Additional information available:

Syntax RulesUsage NotesExample

  o  You must terminate the DEFINE statement  and  the  END  statement
     with periods.

  o  The path name can be a full or  a  relative  path  specification.
     The  last  given  name in the path name is the name you assign to
     the new record definition.  You can specify an  absolute  version
     number with the path name.

  o  The given name of the new record definition is a string of up  to
     31  characters  from  the  set  A-Z,  0-9,  _,  and $.  The first
     character must be a letter from A-Z, and the last character  must
     not  be  _ or $.  You can use 8-bit alphabetic characters in path
     names.

  o  Use the optional DESCRIPTION clause to include text in the CDD to
     document  the  record  definition.   You  must  use  the  keyword
     DESCRIPTION and the text delimiters /* and */ to  insert  one  or
     more  lines  of  text  describing  the  record  into  the  DEFINE
     statement.

  o  The field description statement defines the  whole  record.   You
     can  use an elementary, STRUCTURE, COPY, VARIANTS, or VARIANTS_OF
     field description statement, but  you  can  use  only  one  field
     description statement to define a record.

     STRUCTURE, COPY, and VARIANT  field  description  statements  are
     themselves   subdivided   and   defined   by   subordinate  field
     description statements.

  o  If you specify a path name in the END clause, it must  match  the
     path name in the corresponding DEFINE clause.

  o  If you specify a given name in the END clause,  it  must  be  the
     given  name  of the new data definition and match the final given
     name in  the  path  specification  of  the  corresponding  DEFINE
     clause.

  o  If the path name contains  dictionary  directories  that  do  not
     exist, the CDD automatically creates them.

  o  It is not recommended that you include passwords  in  the  DEFINE
     statement path name, because passwords included in the source can
     be displayed with the DMU LIST/ITEM=SOURCE command.  Instead, use
     the  /PATH  qualifier  with  the compile command if passwords are
     required in the path name.

  o  The only way to assign a password to a record  definition  is  to
     use  the  DMU  SET  PROTECTION or DMU SET PROTECTION/EDIT command
     after you compile the source file.  You cannot assign a  password
     to a record definition by including it in the DEFINE statement.

 The  following  record  description  defines  the  dictionary  object
 SALARY_RECORD  in  the  sample  CDD hierarchy.  The field description
 statement for the record is SALARY STRUCTURE.

 DEFINE RECORD CDD$TOP.PERSONNEL.SERVICE.SALARY_RECORD.
     SALARY STRUCTURE.
         EMPLOYEE_ID          DATATYPE IS UNSIGNED NUMERIC
                              SIZE IS 9 DIGITS.
         PAY STRUCTURE.
             JOB_CLASS        DATATYPE IS UNSIGNED NUMERIC
                              SIZE IS 3 DIGITS.
             INCR_LEVEL       DATATYPE IS UNSIGNED NUMERIC
                              SIZE IS 1 DIGIT.
             WEEKLY_SALARY    DATATYPE IS UNSIGNED NUMERIC
                              SIZE IS 6 DIGITS 2 FRACTIONS.
         END PAY STRUCTURE.
     END SALARY STRUCTURE.
 END SALARY_RECORD RECORD.

 A signed integer specifying the number of places to shift the decimal
 point when the field is evaluated.  A negative scale_factor indicates
 a shift to the left, and a positive scale_factor indicates a shift to
 the right.  Scale_factor must be in the range -128 to 127.

 An unsigned integer indicating the number of characters in the field.

 Represents text in strings of contiguous 8-bit bytes.

 An unsigned integer indicating the number of bytes in the field.

 Specifies the upper bound of a subscript in an ARRAY field  attribute
 clause.  Its value must be greater than or equal to that of the lower
 bound.

 Defines two or more logical views of the same  portion  of  a  record
 definition.   The  VARIANTS  statement  functions  like the REDEFINES
 clause in VAX COBOL and VAX DATATRIEVE.

Additional information available:

ParameterSyntax RuleUsage NotesExample

 variant

      A definition of the field characteristics for  each  subordinate
      field of each VARIANT.

 The VARIANTS, VARIANT, and END statements all must end with periods.

  o  Each VARIANT begins on the same byte in the  record,  subject  to
     individual  alignment options.  The length of the longest VARIANT
     in the collection determines the overall length of  the  VARIANTS
     field description.

  o  Be sure that the VARIANTS collection you define conforms  to  the
     requirements  of  the  language  or  language processor that will
     access the record definition.

  o  VAX DATATRIEVE requires each VARIANT to contain a STRUCTURE field
     description statement at the top of the VARIANT.

 The following example contains three VARIANT  logical  views  of  the
 same record.  In an application program, you can refer to the logical
 view IN_STOCK, BACK_ORDER, or OUT_OF_STOCK depending on how you  want
 to interpret the STOCK field.

 STOCK STRUCTURE.
     VARIANTS.
         VARIANT.
             IN_STOCK STRUCTURE.
                 PRODUCT_NO     DATATYPE IS TEXT
                                SIZE IS 8 CHARACTERS.
                 DATE_ORDERED   DATATYPE IS DATE.
                 STATUS_CODE    DATATYPE IS BYTE.
                 QUANTITY       DATATYPE IS LONGWORD
                                ALIGNED ON LONGWORD.
                 LOCATION       ARRAY 1:4
                                DATATYPE IS TEXT
                                SIZE IS 30 CHARACTERS.
                 UNIT_PRICE     DATATYPE IS LONGWORD SCALE -2.
             END IN_STOCK STRUCTURE.
         END VARIANT.
         VARIANT.
             BACK_ORDER STRUCTURE.
                 PRODUCT_NO     DATATYPE IS TEXT
                                SIZE IS 8 CHARACTERS.
                 DATE_ORDERED   DATATYPE IS DATE.
                 STATUS_CODE    DATATYPE IS BYTE.
                 QUANTITY       DATATYPE IS LONGWORD
                                ALIGNED ON LONGWORD.
                 SUPPLIER       ARRAY 1:4
                                DATATYPE IS TEXT
                                SIZE IS 30 CHARACTERS.
                 UNIT_PRICE     DATATYPE IS LONGWORD
                                SCALE -2.
             END BACK_ORDER STRUCTURE.
         END VARIANT.
         VARIANT.
             OUT_OF_STOCK STRUCTURE.
                 PRODUCT_NO         DATATYPE IS TEXT
                                    SIZE IS 8 CHARACTERS.
                 DATE_LAST_SOLD     DATATYPE IS DATE.
             END OUT_OF_STOCK STRUCTURE.
         END VARIANT.
     END VARIANTS.
 END STOCK STRUCTURE.

 Names a tag variable whose value at run time determines which of  the
 VARIANT fields is the current VARIANT.

Additional information available:

ParametersSyntax RulesUsage NotesExample

 var_field_name

      The tag variable field, whose value determines the selection  of
      the current VARIANT at run time.

 low_value, high_value

     Values to be compared to the value of the  tag  variable  at  run
     time to determine the current VARIANT.


 field description statement

      A definition of the field characteristics  for  the  subordinate
      fields of each VARIANT.

  o  The VARIANT field description block can be repeated.

  o  The tag variable must be an elementary field fixed in the  record
     and  not part of an array, and it must precede the VARIANTS field
     description statement.  It cannot be an unnamed field.

  o  You must fully qualify the tag variable's field name if it is not
     unique  within the record.  A fully qualified field name consists
     of an elementary field name preceded by the  field  names  of  as
     many  of  its  direct  ancestors  as  you  need  to  specify  the
     elementary field uniquely.  You must separate each element  of  a
     fully qualified field name from the next with a period.

     Once the elementary field name is identified  uniquely,  you  can
     omit  any  remaining ancestors' field names.  Furthermore, if the
     tag variable's name is not unique within the record, none of  the
     ancestors in its fully qualified name can be unnamed fields.

  o  You must include a tag value clause for VARIANT.

  o  At run time, the values (low_value and high_value...) you specify
     are  compared  to  the value of the tag variable to determine the
     current VARIANT.

  o  Tag  value  specifications  must   conform   to   the   following
     conditions:

      -  The tag values low_value, high_value...  must be legal values
         as  defined by the tag variable's data type.  For example, if
         the tag variable is a TEXT field,  the  tag  values  must  be
         quoted literals.

      -  The values of high_value must be greater than or equal to the
         values  of  low_value  in  the collating sequence of the data
         type.

      -  The range of values in one VARIANT must not overlap the range
         of values in any other VARIANT.


  o  The compiler ignores commas, but you can use  them  to  make  tag
     value specifications easier to read.

  o  The CDDL compiler does not check that the tag values you  specify
     are legal.  If you specify invalid values, you will receive error
     messages when you refer to the VARIANT field in an application.

  o  The tag value clause specifies a value or set of values  for  the
     tag variable in each VARIANT of a VARIANTS field collection.  The
     tag variable can then be used at run time  to  find  the  current
     VARIANT.

  o  Languages that do not support tag variables ignore the tag  value
     clause.   For  more information on language support of CDD record
     definitions, refer to the documentation for the language you  are
     using.

  o  Each variant begins on the same byte in the  record,  subject  to
     individual  alignment options.  The length of the longest VARIANT
     in the collection determines the overall length of  the  VARIANTS
     field description.

  o  Be sure that the VARIANT collection you define  conforms  to  the
     requirements  of  the  language  or  language processor that will
     access the record definition.

  o  VAX DATATRIEVE requires each VARIANT to contain a STRUCTURE field
     description statement at the top of the VARIANT.

 In the following example, RECORD_IDENTIFIER is the tag variable.  The
 value  of  RECORD_IDENTIFIER  at run time determines which VARIANT is
 current according to the translation table in the descriptive text.

 STOCK STRUCTURE.
     /* RECORD_IDENTIFIER determines field type:
          S --> In-stock record.
          B --> Back-order record.
          O --> Out-of-stock record. */
     RECORD_IDENTIFIER          DATATYPE IS TEXT
                                SIZE IS 1 CHARACTER.
     VARIANTS OF RECORD_IDENTIFIER.
         VARIANT VALUE IS "S".
             IN_STOCK STRUCTURE.
                 PRODUCT_NO     DATATYPE IS TEXT
                                SIZE IS 8 CHARACTERS.
                 DATE_ORDERED   DATATYPE IS DATE.
                 STATUS_CODE    DATATYPE IS BYTE.
                 QUANTITY       DATATYPE IS LONGWORD
                                ALIGNED ON LONGWORD.
                 LOCATION       ARRAY 1:4
                                DATATYPE IS TEXT
                                SIZE IS 30 CHARACTERS.
                 UNIT_PRICE     DATATYPE IS LONGWORD SCALE -2.
             END IN_STOCK STRUCTURE.
         END VARIANT.
         VARIANT VALUE IS "B".
             BACK_ORDER STRUCTURE.
                 PRODUCT_NO     DATATYPE IS TEXT
                                SIZE IS 8 CHARACTERS.
                 DATE_ORDERED   DATATYPE IS DATE.
                 STATUS_CODE    DATATYPE IS BYTE.
                 QUANTITY       DATATYPE IS LONGWORD
                                ALIGNED ON LONGWORD.
                 SUPPLIER       ARRAY 1:4
                                DATATYPE IS TEXT
                                SIZE IS 30 CHARACTERS.
                 UNIT_PRICE     DATATYPE IS LONGWORD
                                SCALE -2.
             END BACK_ORDER STRUCTURE.
         END VARIANT.

         VARIANT VALUE IS "O".
             OUT_OF_STOCK STRUCTURE.
                 PRODUCT_NO         DATATYPE IS TEXT
                                    SIZE IS 8 CHARACTERS.
                 DATE_LAST_SOLD     DATATYPE IS DATE.
             END OUT_OF_STOCK STRUCTURE.
         END VARIANT.
     END VARIANTS.
 END STOCK STRUCTURE.

 An unsigned integer indicating the number of characters in the field.