Re: Extra record on iSeries to PC FTP

[CRPence <crp@xxxxxxxxxxxxxxxxxxxx>]

  The DMP command can be used to review [DSPSPLF QPSRVDMP 
SPLNBR(*LAST)] for the last two bytes of the file data, to verify they 
are indeed 0x0D0A

  Whether zero bytes following a CRLF constitutes an /extra record/ is 
unfortunately open to interpretation.  If the receiving application 
insists that zero bytes of data following the last row terminated by 
CRLF [i.e. End Of Record delimiter for a stream of text data is 0x0D0A] 
constitutes a new row, then it would seem appropriate to concur with 
them, and simply tell them that the /extra record/ is there; i.e. by 
counting the same way they do, when generating the control file. 
Although as an effective null record [i.e. no data], it seems unlikely a 
database as recipient will deal well with that -- maybe create a row 
where all fields are the NULL value.? Yikes!  I would expect that the 
EOF would be noted, and thus the import would understand there are no 
more rows.

  Some people will insist that as an EOR delimiter, the CRLF must exist 
on the last record, just as in every previous record.  Others insist 
there is no legitimate reason to include an EOR delimiter prior to an 
EOF marker nor as the last two bytes; in fact claim that it is an error. 
 I know this was discussed back in the day, and I recall only that some 
chose to rewrite the file dropping the last two bytes, or otherwise with 
the same outcome, using facilities to truncate the last two bytes [if 
they were x'0D0A'].  Seems it would be better if the recipient can just 
understand what is meant by <data><EOR><EOF>; where <EOF> is not to 
imply an actual delimiter\marker, as it could be implied by zero bytes 
remaining.

  CPYTOSTMF [and CPYTOIMPF, at least in its pre-V5R3 design which, in 
its implementation, used CPYTOSTMF] when copying from a database file 
will terminate the line after the last non-blank character in a record 
according to the ENDLINFMT() value specified [on CPYTOIMPF, specified 
with RCDDLM() parameter].  EDTF seems to mutter if the delimiter is 
missing, and will also add the EOR delimiter for the last record when 
the file is changed, but that interface does not present any /extra/ 
/line/ of text.

Regards, Chuck