On 11-Feb-2014 07:04 -0800, Paul Bailey wrote:
I am using RDi 9.0.1 on v6r1. When I enter a CPYF command into a
CLLE source member I usually have no problem.
When I enter the following into CLLE source, Rational complains with
error "EVFCL0071: Parameter INCREL required.":
CPYF FROMFILE(FILE1) TOFILE(FILE2) MBROPT(*ADD) +
INCREL((*IF FLD1 *EQ ' '))
If FLD1 is not a 1-byte char field, pad the literal value [specified
as the fourth element of the first list-element in the INCREL] to the
full length of the field FL1 to ensure including records whereby the
value of FLD1 is all blanks if that is the intention, versus including
records whereby just the first byte of FLD1 is a blank. As coded, the
latter is the effect of the request.
I can't prompt (F4) the line. If I remove the INCREL clause, I can
then prompt the line, but I can not "add" the INCREL option with ' '
as the value.
As a workaround I am declaring a variable as ' ' and using the
variable instead of the quoted string, but all of this works fine in
SEU.
<<SNIP>>
As a circumvention for the prompting issue, use of a variable name is
acceptable. Be aware however, that care must be taken...
If that circumvention is carried\compiled into the actual code,
nuances of using a variable in the actual compiled code to provide a
comparison value [in any of the CPYF parameters] may yield unexpected
results [for anyone unaware of the how to overcome those nuances]. To
effect a comparison against the full length of the field, the variable
typically would be declared as one byte longer than the field, with the
last byte assigned to any non-blank value, to ensure any blank-padding
is accounted for. I suggested in the following archived message, that
the INCREL is not eligible for that circumvention of using a larger
variable, per error msg CPF2843, so I noted potential alternatives
[other than building the command string such that the variable becomes
the delimited literal value]:
<
http://archive.midrange.com/midrange-l/201107/msg00408.html> I doubt
that restriction for the longer length variable was ever lifted; FWiW,
verified to exist still on v5r3. A number of discussions in the
archives exist with tokens CPYF and INCREL; the site seems not to be
responding presently, so I did not look for and add any other good links
for reference.
The issue is that... A trimmed [non-zero] length of the compare-value
from a variable, is the length of data from the field that is compared
against; this is similar for a literal compare value, but a padded
literal value is seen by the CPP as the full length versus the effective
trimmed length. Thus having compared a variable having the value of
blanks against a field, yields a one-byte blank compare against the
first one byte of data in that field; i.e. if FLD1 in the scenario is a
CHAR(1), then there is no issue, however if the FLD1 is
CHAR(2_or_greater) then:
A row with FLD1=' LeftPad' will be included in the results for
INCREL((*IF FLD1 *EQ &char10)) if the value of &char10 *eq ' '.
midrange.com
