Re: Corrupt Data

On 28-Mar-2014 09:51 -0700, rob@xxxxxxxxx wrote:

On 28-Mar-2014 09:26 -0700, Srinivas Boggula wrote:

We have converted some data from text file to PF file. The text
file has lines separated by CRLF. This PF file when read in V5R4M0
doesn't create any problem.
But when being read in V7R1M0 this CRLF is being consider as
character. <<SNIP>>

I suspected you used CPYFRMIMPF to do this conversion. IBM changes
that command with every release of the OS. Please read BOTH the 6.1
Memo To Users and the 7.1 Memo To Users. I am sure that command will
be mentioned in it. This constant changing of the behavior of this
command has led me to stop using it. <<SNIP>>

Anyone desiring the effects similar to the older defect-laden and performance impaired import feature, perhaps because they had became dependent on the defects [e.g. EOR control character(s) embedded in strings incorrectly effecting the breaking of the data into two database records, thus causing the remaining string data to fill the data of the first column in the next row], then there is the option of preceding the import request with CPYFRMSTMF into a source physical file of a sufficiently accommodating record length, and then modifying their import request to reference that SRC-PF member as input instead of referring to the stream file. Or, there is [at least there was] a data area that can be defined to force the v5r3 implementation of the function to be invoked; thus willfully reintroduce all of the defects from the past [which includes the use of CPYFRMSTMF by the import feature, as its first phase, and that use implicitly and incorrectly causes the string data with embedded EOR to break into a new record].