When assigning match field values (M1 through M9) to fields on the input
specifications in positions 65 and 66, consider the following:
Sequence
checking is done for all record types with match field specifications. All
match fields must be in the same order, either all ascending or all descending.
The contents of the fields to which M1 through M9 are assigned are checked
for correct sequence. An error in sequence causes the RPG IV exception/error
handling routine to receive control. When the program continues processing,
the next record from the same file is read.
Not all files used in the program must have match fields. Not all record
types within one file must have match fields either. However, at least one
record type from two files must have match fields if files are ever to be
matched.
The same match field values must be specified for all record types that
are used in matching. See Figure 44.
Date, time, and timestamp match fields with the same match field values
(M1 through M9) must be the same type (for example, all date) but can be different
formats.
All character, graphic, or numeric match fields with the same match field
values (M1 through M9) should be the same length and type. If the match field
contains packed data, the zoned decimal length (two times packed length -
1) is used as the length of the match field. It is valid to match a packed
field in one record against a zoned decimal field in another if the digit
lengths are identical. The length must always be odd because the length of
a packed field is always odd.
Record positions of different match fields can overlap, but the total
length of all fields must not exceed 256 characters.
If more than one match field is specified for a record type, all the fields
are combined and treated as one continuous field (see Figure 44).
The fields are combined according to descending sequence (M9 to M1) of matching
field values.
Match fields values cannot be repeated in a record.
All match fields given the same matching field value (M1 through M9) are
considered numeric if any one of the match fields is described as numeric.
When numeric fields having decimal positions are matched, they are treated
as if they had no decimal position. For instance 3.46 is considered equal
to 346.
Only the digit portions of numeric match fields are compared. Even though
a field is negative, it is considered to be positive because the sign of the
numeric field is ignored. Therefore, a -5 matches a +5.
Date and time fields are converted to *ISO format for comparisons
Graphic data is compared hexadecimally
Whenever more than one matching field value is used, all match fields
must match before the MR indicator is set on. For example, if match field
values M1, M2, and M3 are specified, all three fields from a primary record
must match all three match fields from a secondary record. A match on only
the fields specified by M1 and M2 fields will not set the MR indicator on
(see Figure 44).
UCS-2 fields cannot be used for matching fields.
Matching fields cannot be used for lookahead fields, and arrays.
Field names are ignored in matching record operations. Therefore, fields
from different record types that are assigned the same match level can have
the same name.
If an alternate collating sequence or a file translation is defined for
the program, character fields are matched according to the alternate sequence
specified.
Null-capable fields, character fields defined with ALTSEQ(*NONE), and
binary, float, integer and unsigned fields (B, F, I, or U in position 36 of
the input specifications) cannot be assigned a match field value.
Match fields that have no field record relation indicator must be described
before those that do. When the field record relation indicator is used with
match fields, the field record relation indicator should be the same as a
record identifying indicator for this file, and the match fields must be grouped
according to the field record relation indicator.
When any match
value (M1 through M9) is specified for a field without a field record relation
indicator, all match values used must be specified once without a field record
relation indicator. If all match fields are not common to all records, a dummy
match field should be used. Field record relation indicators are invalid for
externally described files. (see Figure 45).
Match fields are independent of control level indicators (L1 through L9).
If multi-file processing is specified and the LR indicator is set on,
the program bypasses the multi-file processing routine.
Figure 44 is
an example of how match fields are specified.
Figure 44. Match Fields in Which All Values Match
*...1....+....2....+....3....+....4....+....5....+....6....+....7...
FFilename++IPEASFRlen+LKlen+AIDevice+.Keywords++++++++++++++++++++++++++++
* The files in this example are externally described (E in position
* 22) and are to be processed by keys (K in position 34).
FMASTER IP E K DISK
FWEEKLY IS E K DISK
*...1....+....2....+....3....+....4....+....5....+....6....+....7...
IRcdname+++....Ri........................................................
I..............Ext-field+..................Field+++++++++L1M1..PlMnZr....
* MASTER FILE
IEMPMAS 01
I EMPLNO M1
I DIVSON M3
I DEPT M2
IDEPTMS 02
I EMPLNO M1
I DEPT M2
I DIVSON M3
* WEEKLY FILE
IWEEKRC 03
I EMPLNO M1
I DIVSON M3
I DEPT M2
Three files are used in matching records. All the files have three match
fields specified, and all use the same values (M1, M2, M3) to indicate which
fields must match. The MR indicator is set on only if all three match fields
in either of the files EMPMAS and DEPTMS are the same as all three fields
from the WEEKRC file.
The three match fields in each file are combined and treated as one match
field organized in the following descending sequence:
DIVSON
M3
DEPT
M2
EMPLNO
M1
The order in which the match fields are specified in the input specifications
does not affect the organization of the match fields.
Figure 45. Match Fields with a Dummy M2 Field
*...1....+....2....+....3....+....4....+....5....+....6....+....7...
IFilename++SqNORiPos1+NCCPos2+NCCPos3+NCC................................
I........................Fmt+SPFrom+To+++DcField+++++++++L1M1FrPlMnZr....
IDISK AB 01 1 C1
I OR 02 1 C2
I OR 03 1 C3
I 1 10 0EMPNO M1
I 11 15 0DUMMY M2
I 11 15 0DEPT M202
I 16 20 0DEPT M203
Three different
record types are found in the input file. All three contain a match field
in positions 1 through 10. Two of them have a second match field. Because
M1 is found on all record types, it can be specified without a field record
relation entry in positions 67 and 68. If one match value (M1 through M9)
is specified without field record relation entries, all match values must
be specified once without field record relation entries. Because the value
M1 is specified without field record relationship, an M2 value must also be
specified once without field record relationship. The M2 field is not on
all record types; therefore a dummy M2 field must be specified next. The
dummy field can be given any unique name, but its specified length must be
equal to the length of the true M2 field. The M2 field is then related to
the record types on which it is found by field record relation entries.
Figure 46. Match Field Specifications for Three Disk Files
*...1....+....2....+....3....+....4....+....5....+....6....+....7...
FFilename++IPEASFRlen+LKlen+AIDevice+.Keywords++++++++++++++++++++++++++++
FPRIMARY IPEA F 64 DISK
FFIRSTSEC IS A F 64 DISK
FSECSEC IS A F 64 DISK
*...1....+....2....+....3....+....4....+....5....+....6....+....7...
IFilename++SqNORiPos1+NCCPos2+NCCPos3+NCC................................
I........................Fmt+SPFrom+To+++DcField+++++++++L1M1FrPlMnZr....
IPRIMARY AA 01 1 CP 2NC
I 2 3 MATCH M1
*
I BB 02 1 CP 2 C
I 2 3 NOM
*
IFIRSTSEC AB 03 1 CS 2NC
I 2 3 MATCH M1
*
I BC 04 1 CS 2 C
I 2 3 NOM
*
ISECSEC AC 05 1 CT 2NC
I 2 3 MATCH M1
*
I BD 06 1 CT 2 C
I 2 3 NOM
