If you have a null-capable keyfield,
you can position to the beginning or end of the file using
the SETLL operation with *START or *END.
If you have a null-capable key field, you can search for records containing null values by specifying an indicator in factor 2 of the KFLD operation and setting that indicator on before the keyed input operation. If you do not want a null key to be selected, you set the indicator off.
If factor 2 of a KFLD operation is not specified,
then the null-key-byte-map information will be set to zero for that key.
When a record format with null-capable key fields is used on a CHAIN, SETLL, READE, or READPE operation, and a %KDS data structure is used to specify the keys, then the null-key-byte-map information will be taken from the null attributes of the subfields in the data structure specified as the argument of %KDS.
When a record format with null-capable key fields is used on a CHAIN, SETLL, READE, or READPE operation, and a list of keyfields is used, then the null-key-byte-map information will be taken from the null attributes of the specified keys.
Figure 1 and Figure 2 illustrate how keyed operations are used to position and retrieve records with null keys.
// Assume File1 below contains a record Rec1 with a composite key
// made up of three key fields: Key1, Key2, and Key3. Key2 and Key3
// are null-capable. Key1 is not null-capable.
// Each key field is two characters long.
*..1....+....2....+....3....+....4....+....5....+....6....+....7....+..
FFilename++IPEASFRlen+LKlen+AIDevice+.Keywords++++++++++++++++++++++++++
FFile1 IF E DISK
// Define two data structures with the keys for the file
// Subfields Key2 and Key3 of both data structures will be
// null-capable.
DName+++++++++++ETDsFrom+++To/L+++IDc.Keywords+++++++++++++++++++++++++
D Keys DS LIKEREC(Rec1 : *KEY)
D OtherKeys DS LIKEDS(keys)
// Define a data structure with the input fields of the file
// Subfields Key2 and Key3 of the data structures will be
// null-capable.
D File1Flds DS LIKEREC(Rec1 : *INPUT)
/free // The null indicator for Keys.Key2 is ON and the
// null indicator for Keys.Key3 is OFF, for the
// SETLL operation below. File1 will be positioned
// at the next record that has a key that is equal
// to or greater than 'AA??CC' (where ?? is used
// in this example to indicate NULL)
// Because %NULLIND(Keys.Key2) is ON, the actual content
// in the search argument Keys.Key2 will be ignored.
// If a record exists in File1 with 'AA' in Key1, a null
// Key2, and 'CC' in Key3, %EQUAL(File1) will be true.
Keys.Key1 = 'AA';
Keys.Key3 = 'CC';
%NULLIND(Keys.Key2) = *ON;
%NULLIND(Keys.Key3) = *OFF;
SETLL %KDS(Keys) Rec1;
// The CHAIN operation below will retrieve a record
// with 'JJ' in Key1, 'KK' in Key2, and a null Key3.
// Since %NULLIND(OtherKeys.Key3) is ON, the value of
// 'XX' in OtherKeys.Key3 will not be used. This means
// that if File1 actually has a record with a key
// 'JJKKXX', that record will not be retrieved.
OtherKeys.Key3 = 'XX';
%NULLIND(Keys.Key3) = *ON;
CHAIN ('JJ' : 'KK' : OtherKeys.Key3) Rec1;
// The CHAIN operation below uses a partial key as the
// search argument. It will retrieve a record with 'NN'
// in Key1, a null key2, and any value including a null
// value in Key3. The record is retrieved into the
// File1Flds data structure, which will cause the
// null flags for File1Flds.Key2 and File1Flds.Key3
// to be changed by the operation (if the CHAIN
// finds a record).
Keys.Key1 = 'NN';
%NULLIND(Keys.Key2) = *ON;
CHAIN %KDS(Keys : 2) Rec1 File1Flds;
* Using the same file as the previous example, define two
* key lists, one containing three keys and one containing
* two keys.
CL0N01Factor1+++++++Opcode(E)+Factor2+++++++Result++++++++Len++D+HiLoEq.
C Full_Kl KLIST
C KFLD Key1
C KFLD *IN02 Key2
C KFLD *IN03 Key3
C Partial_Kl KLIST
C KFLD Key1
C KFLD *IN05 Key2
*
* *IN02 is ON and *IN03 is OFF for the SETLL operation below.
* File1 will be positioned at the next record that has a key
* that is equal to or greater than 'AA??CC' (where ?? is used
* in this example to indicate NULL)
*
* Because *IN02 is ON, the actual content in the search argument
* for Key2 will be ignored.
*
* If a record exists in File1 with 'AA' in Key1, a null Key2, and
* 'CC' in Key3, indicator 90 (the Eq indicator) will be set ON.
*
C MOVE 'AA' Key1
C MOVE 'CC' Key3
C EVAL *IN02 = '1'
C EVAL *IN03 = '0'
C Full_Kl SETLL Rec1 90 *
* The CHAIN operation below will retrieve a record with 'JJ' in Key1,
* 'KK' in Key2, and a null Key3. Again, because *IN03 is ON, even
* if the programmer had moved some value (say 'XX') into the search
* argument for Key3, 'XX' will not be used. This means if File1
* actually has a record with a key 'JJKKXX', that record will not
* be retrieved.
*
C MOVE 'JJ' Key1
C MOVE 'KK' Key2
C EVAL *IN02 = '0'
C EVAL *IN03 = '1'
C Full_Kl CHAIN Rec1 80
*
* The CHAIN operation below uses a partial key as the search argument.
* It will retrieve a record with 'NN' in Key1, a null key2, and any
* value including a null value in Key3.
*
* In the database, the NULL value occupies the highest position in
* the collating sequence. Assume the keys in File1 are in ascending
* sequence. If File1 has a record with 'NN??xx' as key (where ??
* means NULL and xx means any value other than NULL), that record
* will be retrieved. If such a record does not exist in File1, but
* File1 has a record with 'NN????' as key, the 'NN????' record will
* be retrieved. The null flags for Key2 and Key3 will be set ON
* as a result.
*
C MOVE 'NN' Key1
C SETON 05
C Partial_Kl CHAIN Rec1 70