MODE-2350 OracleDdlParser update for multi-op ALTER TABLE

[freakolowsky]

Method parseAlterTableStatement now supports basic chained ADD and/or
MODIFY commands. The existing functionality should be the same (ran
tests without issues).

Will do some more coding later to add complete column definitions as
ALTER TABLE MODIFY currently only detects changes that include datatype
modifications.

• modified parseAlterTableStatement
• modified parseColumns and overloaded for compatibility
• created Oracle local parseColumnDefinition
• hopefuly not broken things (... much)

[hchiorean]

looks fine, provided there are no regressions.
@rhauch: I guess we'd a new JIRA enhancement for this ?

[rhauch]

Yes, we need a new enhancement JIRA that adequately explains the changes in terms of the DDL expressions to support. Then please update the commit message (using git amend is easiest) to include the JIRA number at the front (e.g., MODE-5678).

It'd probably be good to add a few more unit tests that cover not just the new capability, but also handle some failure cases, such as when the new expressions are formed incorrectly.

[hchiorean]

@freakolowsky: can you pls open the enhancement JIRA, tx.

[freakolowsky]

Like this? ... sry, but i'm new in this part of town :D

[hchiorean]

@freakolowsky: yes, that's pretty close. tx.

[freakolowsky]

oh yes ... @rhauch incorrectly formed expression should already be handled by existing tests, this mod only adds support for ADD/MODIFY repeats

@hchiorean: what am i missing? if i'll poking around this code i don't want to piss pplz off on the first try :D

[hchiorean]

@freakolowsky: in the context of the changes (just supporting multiple ADD and MODIFY statements within ALTER TABLE) not sure what rhauch meant by

handle some failure cases, such as when the new expressions are formed incorrectly.

@rhauch ^^ ?

[rhauch]

@hchiorean, @freakolowsky: My comment about testing improper expressions is simply to verify that the parser correctly generates a useful exception/problem when an expression is poorly formed. Doesn't need to be anything fancy or really that many. IOW, the repeating ADD and MODIFY statements are what's new, so try to test a couple of expressions that use them incorrectly just to make sure that a decent error/problem is reported.

[rhauch]

@freakolowsky: Overall, this is stellar. Thank you for the contribution! Yeah, there are a couple of minor things we wanted changed, mostly just to follow our change process (file an issue, make sure the PR references the issue, make sure the issue references the PR, etc.).

You also wrote:

if i'll poking around this code i don't want to piss pplz off on the first try :D

No worries there! In fact, we're very grateful for the improvement! Keep up the good work!

[freakolowsky]

@rhauch np ... this chunk of modeshape project saved me from having to write a DDL parser from scratch which i need for another project

[hchiorean]

@freakolowsky: based on the previous comments, I'd say it's safe to amend this & add a couple of more tests with invalid ADD & MODIFY sequences and after that I'll merge this.

[freakolowsky]

Will try to do it over the weekend or next week ...

[hchiorean]

@freakolowsky: any update on this ?

[rhauch]

@freakolowsky Do you think you might update the PR soon?

[freakolowsky]

sry 4 the delay ... had some personal issues (father got cancer) and had to spend a lot of time out of the office. Will update asap. Don't give up on me yet :D

[rhauch]

@freakolowsky, no worries. Thanks for the reply, and I'm very sorry to hear about your father. We can keep this PR open until you have a chance to make the changes.

[hchiorean]

superseded by #1383