The following list describes the enhancements made to ILE COBOL in 7.2:
- TIMESTAMP support of 0 to 12 fractional seconds
A timestamp
item can now have between 0 and 12 fractional seconds.
- The following intrinsic functions will now allow PICOSECONDS as
a duration when specified for a timestamp item: ADD-DURATION, EXTRACT-DATE-TIME,
FIND-DURATION, and SUBTRACT-DURATION.
- The SIZE keyword is allowed with FORMAT TIMESTAMP. The size can
be 19 indicating zero fractional seconds or a value between 21 and
32 indicating between 1 and 12 fractional seconds.
- XML PARSE now has the capability to parse XML files that are greater
than 16MB in size, provided that no individual document piece passed
to the processing procedure is greater than 16MB. The following new
XML-CODE values are associated with this change:
- XML-CODE 62 indicates that the XML document exceeds 16,000,000
bytes.
- XML-CODE 170 indicates that an XML event exceeds 16,000,000 bytes.
- PCML generation
- PCML generation provides improved OCCURS DEPENDING ON array handling
with the addition of a new "init" keyword that will be set to the
maximum size of the array.
- PCML generation provides automatic data-item naming in generated
PCML for filler data items and unnamed items in a data structure,
helping to enable web services to use generated PCML without first
modifying it. The names for these data items will be _filler_1, _filler_2,
and so on.
- National (Unicode) enhancements
- Numeric national datatype is supported
- A numeric literal can be specified on the VALUE clause for a numeric national
data item
- The figurative constant ZERO/ZEROS/ZEROES represents one or more national
zero digits when used with national data items
- National 'N' literals are supported when new PROCESS option NATIONALPICNLIT
is specified
- The accuracy of numeric intrinsic functions NUMVAL and NUMVAL-C
increases to 31 digits with compiler option ARITHMETIC(*EXTEND31)
or PROCESS option EXTEND31.
- ARITHMETIC parameter for CRTBNDCBL / CRTCBLMOD:
New *EXTEND31FULL
option value provides the following features:
- The accuracy of the following numeric intrinsic functions increases
from floating-point accuracy of up to 15 digits to decimal floating-point
accuracy of up to 34 digits: ANNUITY, MEAN, MEDIAN, MIDRANGE, NUMVAL,
NUMVAL-C, PRESENT-VALUE, and VARIANCE.
- The intermediate result of a fixed-point arithmetic expression
can be up to 34 digits and numeric literals may have a maximum length
of 34 digits.
- New PROCESS statement options:
- NOCHGFLTRND / ALWCHGFLTRND
Specifies whether or not COBOL will
use the floating point rounding mode computational attribute specified
by MI instruction SETCA. SETCA allows you to set the rounding mode
of the result of a floating-point calculation to either round or truncate.
- NATIONALPICNLIT
Enables N" and N' as the opening
delimiter for a national literal and enables elementary data items
defined using the picture symbol N to have an implied USAGE NATIONAL
clause.
- EXTEND31FULL
Note: There may be screen captures in this guide that contain obsolete
references to iSeries®.