JSqlParser
JSqlParser copied to clipboard
JSQLParser
Closes #1604, added simple OVERLAPS support
- Using the
ExpressionListapproach as discussed in the other PR - No renames of existing classes
- Moved to
SQLCondition
Rob,
there was an unrelated PR merged last night, which breaks the tests. I have provided a fix already, but we will need to wait for it to get merged first.
Until this happens, all our PRs will fail to built. So please a bit patience right now. Thanks, its not in my hands.
I noticed. I spent some time trying to figure out why the AlterTest was failing, until I decided to check if the main branch was the issue, and it was.
(You're still welcome to provide code-review-feedback already of course)
I definitely will, just wait please until Master builds again. It wasn't me but it affects my work as well.
It's a thing of beauty now and well executed!
One recommendation: You may want to add <OVERLAPS> to the RelObjectWithoutValue Production in order to allow it as Column, Table, Alias, Function. (I still hope that my Keywords PR will be allowed to do that automatically eventually.)
One question: Are those ExpressionList() really of a variable length? Or are they limited to strictly 2? Should we want to enforce it?
Well done and thank you for your contribution!
Thanks :)
You meant: RelObjectNameWithoutValue right? I'm not really sure if that is correct/something desired? At least PostgreSQL does not allow it to be used as Table without quoting it.
So CREATE TABLE OVERLAPS(A TEXT); gives a syntax error; CREATE TABLE "OVERLAPS"(A TEXT); is allowed.
If there is consensus that it should be allowed, I'm happy to add it! (or maybe, after the keyword PR / discussion is settled?)
About the ExpressionList: I tried to stick to the sql2003 standard. I'm not too confident about the different DBMS handling of OVERLAPS. The standard allows a list of size 2+, or, when prefixed with ROW, a list of size 1+. The latter I did not implement as I tried to follow the mantra 'we will implement it when needed' :)
WDYT?
You meant:
RelObjectNameWithoutValueright? Yes.
I'm not really sure if that is correct/something desired? At least PostgreSQL does not allow it to be used as Table without quoting it.
So
CREATE TABLE OVERLAPS(A TEXT);gives a syntax error;CREATE TABLE "OVERLAPS"(A TEXT);is allowed.
As long as it is not explicitly defined by the SQL Standard as a Reserved Keyword we should try to allow it as a keyword.
If there is consensus that it should be allowed, I'm happy to add it! (or maybe, after the keyword PR / discussion is settled?)
I have no particular interest in this one Keyword. But I wanted to point out the mechanism -- for your next great contribution.
About the
ExpressionList: I tried to stick to the sql2003 standard. I'm not too confident about the different DBMS handling ofOVERLAPS. The standard allows a list of size 2+,
Perfect then, this was all I wanted to know. (I have never used OVERLAPS and I am still very stuck with Oracle and SQL 2002, sorry.)
I have no particular interest in this one Keyword. But I wanted to point out the mechanism -- for your next great contribution.
I will leave this for now, if there arises an issue/request for it, we can easily add it.
Perfect then, this was all I wanted to know
Thanks for the question; I'll leave a comment in the grammar pointing this out, that's useful for other ppl. that might wonder why this has been done this way.
Yes, it certainly can. However only @wumpz can merge.On 16 Aug 2022 08:50, Rob Audenaerde @.***> wrote: @manticore-projects can this be pulled now? :)
—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>
You are my hero, Sir!
Aug 18, 2022 8:10:28 PM net.sf.jsqlparser.statement.select.SpecialOracleTest recordSuccessOnSourceFile
INFO: NEW SUCCESS: interval05.sql
Aug 18, 2022 8:10:29 PM net.sf.jsqlparser.statement.select.SpecialOracleTest testAllSqlsParseDeparse
INFO: tested 276 files. got 192 correct parse results, expected 191
You increased the Oracle Coverage/Compliance by 1 test -- and nobody has noticed.
Please go to src/test/java/net/sf/jsqlparser/statement/select/SpecialOracleTest.java and register your succeeding test as I do not want to steal the merit.
Thank you again for the contribution and cheers!
I really don't mind anybody 'stealing' the merit, so go ahead :)
(I just care about having a great tool to help me in my SQL conquests, and am happy to be able to contribute where/when possible)