Re: Copying a CSV file to iSeries as bar delim flat file

On 11-Jan-2012 09:01 , sjl wrote:

IMHO, in the case of the CPYTOIMPF/CPYFRMIMPF command
"enhancements", IBM forgot their long-standing promise to the AS/400
- iSeries - System i - Power community, which was always represented
to me as:

'We will preserve your investment in your existing software'

Which I always took to mean that for the most part, you don't
experience the pain that software developers on other platforms
(particularly Windows) experience when migrating your software to a
new release/version of the O/S.

i.e., you don't have to rewrite your code every time IBM comes out
with a new release.

Yes, but only up to a point. And most notably, for compiled HLL code. Products and features are dropped, implementations change, and corrections and even incompatible changes are made.

A poor performing utility which was completely flawed in several aspects and identified as such in many APARs, is not something that would have best remained unchanged. The feature was rewritten specifically because it was so flawed; the most basic rule of being able to import what was just exported was even a problem with that utility, before the rewrite. I agree the transition was more harsh than preferable [i.e. could have been effected with fewer impacts], but a data area was enabled to give the same old incorrect results and poor performance as in prior releases, to anyone who desired. Both the origin of the code for the feature and the intent for and expectation how the import\export would be used [impromptu versus embedded into applications] resulted in [rightly, IMO] less concern for an emphasis on compatibility. Besides, for the vast majority of import problems experienced\reported, resolution could be as simple as inserting a CPYFRMSTMF before the CPYFRMIMPF; i.e. to insert what the updated CPYFRMIMPF was no longer doing [as a change to prevent an unnecessary intermediate copy], and then the updated import utility worked much like before, except with a number of corrections and enhancements available in parsing the input data.

As to changes which are known in advance or later identified to cause difficulties due to an upgrade, IBM offers the "Memo to Users" or MTU. Those affected may indeed have to make updates and\or rewrite some code :-( Unfortunately some defects must be corrected and those who relied on them must also make changes, as well for other nuanced or blatant changes for which accommodations will not be made then those using the changed features also must change. There are limits to what can be preserved; esp. true when the "existing software" was coded to depend on defects, which was often the case for the output from the older database import\export utility.

Regards, Chuck