Re: (Date/Time fields in DB file)

Hello Rob,

Am 11.08.2023 um 13:26 schrieb Rob Berendt <robertowenberendt@xxxxxxxxx>:

I differ. Not that I can recall the last time I fired up WRKMBRPDM, and STRPDM has been years, but PDM allows you to create user defined options you could bury RUNSQLSTM in.

Yap, I'm aware.

I'll grant you that I'm not sure that V4R5 supports CREATE OR REPLACE TABLE.

Well, it's not only that. What I missed most when migrating stuff from Linux/MySQL to OS/400 was an auto-incrementing integer field. That's only supported with newer OS/400 releases and can only be defined with SQL.

Meanwhile I have my copy-paste code templates for that, but it was quite a difference to grok. Surely there might be more examples about "not with V4R5", but I can't recall (to me) important examples.

I don't complain. That's the price I pay for deliberately using an old, comparably slow system for hobbyist purposes. But it runs a really "cool" OS I greatly enjoy working with, and with comparably little power consumption. And it's overall definitely way less boring than Linux on each and any arbitrary hardware platform.

And DDL without CREATE OR REPLACE has a different paradigm where you often just use ALTER TABLE vs recreating it.

I understand. Allow me to clarify: Seeing the CREATE TABLE definition with all those ALTER TABLE statements already integrated helps me to quickly grok a table's layout. That stems from my earlier habit to use `mysqldump --no-data` to see the complete and current definition of a table.

That's what I love about DDS: Quick glance, profound overview.

If you want the raw DDL recreated as you're not sure if the existing DDL is in sync with reality there are APIs to generate the DDL from the table.

And exactly that would be a big step towards adhering to your initial "quit that" advice. :-)

Funny you mention CHGPF with DDS. How many still run into people who weren't aware that has been supported for decades and they still run into people who create a new table and copy everything into it from the old version?

Well, the command naming IBM came up with was – and is — extremely helpful in guessing a command's name when you already know what you want to "do". That's how I learned about CHGPF. Especially great is that it also handles all the background work with dependent LFs, and journalling.

:wq! PoC