Re: Making the most, or over-thinking it - SQL for uploads

On 30 August 2016 at 18:56, John Yeung <gallium.arsenide@xxxxxxxxx> wrote:

What most people are saying, and I have to agree, is that transferring
the data from the (entire) text file into a database is ultimately not
a big deal. It's a "separate step" but that step is not a very
expensive one, and if you are building a recurring process, it can be
automated so that you don't even see it happening.

Well thread drift has pretty much removed RDi as the subject so this
will be my last post on it here. Many people over on MIDRANGE-L have
dealt with this problem; follow-ups will probably get definitive
answers there quite quickly.

What I do is share an IFS directory with Netserver. The Windows PC
user simply maps a drive letter to that directory and when she does a
'Save As...' it saves the file directly to the IFS, where any number
of tools can deal with it.

An alternative is to use QNTC and share the Windows PC directory, and
have IBM i 'reach out' and treat the Windows PC as though it were an
IFS directory.

If the text file comes from a trading partner, IBM i can get the file
with a utility like Scott Klement's HTTPAPI, or with the new HTTP
services in DB2 SQL :-)

Point being that there doesn't need to be a separate, manual
'transfer' step in this process.
--buck