[antlr-interest] Re: Very weird SQL parser failures - look-ahead? - FIXED

[Lubos Vnuk]

Hi Toly,

> Is there a simple way in Antlr to specify pairs of optional 
> matching parenthesis (I mean apart from brute-force when you mark 
> them explicitly)?
You might want to have a look at my SQL-92 grammar at ANTLR website. 
Generally speaking, this is typically done by the recursive 
expression rules.

Also, in order to get a quality grammar for a SQL dialect, I 
recommend starting off by transcribing the syntax diagrams that have 
usually been published in the DBMS documentation. They have BNF-close 
value and provide a great implementation guidance.

HTH,
Lubos.


--- In antlr-interest at yahoogroups.com, "anatoli_krassavine" 
<toly at i...> wrote:
> Hello Monty,
> 
> Great thanks to you and Bharath for comments - it pushed me in 
> correct direction. There was a stupid grammar mistake which only 
> showed when I used -traceParser.
> 
> I will definitely take a look at NIST SQL-92 test suite as you 
> adviced.
> 
> Since you just finished a SQL parser of your own, could I ask a 
quick 
> question. Is there a simple way in Antlr to specify pairs of 
optional 
> matching parenthesis (I mean apart from brute-force when you mark 
> them explicitly)?
> 
> Sincerely,
>    Toly
> 
> --- In antlr-interest at yahoogroups.com, Monty Zukowski <monty at c...> 
> wrote:
> > Run antlr.Tool with -traceParser, that may help you figure out 
> where it 
> > is going wrong.
> > 
> > Also, I just finished a SQL parser on contract.  I recommend the 
> NIST 
> > SQL-92 test suite to ferret out grammatical problems such as 
these.
> > 
> > I am also available by the hour to debug things like this.  
> > Unfortunately I don't have so much "free" time to solve these 
> problems 
> > for fun any more.
> > 
> > Monty Zukowski
> > 
> > ANTLR & Java Consultant -- http://www.codetransform.com
> > ANSI C/GCC transformation toolkit -- 
> > http://www.codetransform.com/gcc.html
> > Embrace the Decay -- 
http://www.codetransform.com/EmbraceDecay.html
> > 
> > 
> > 
> > 
> > On Jun 1, 2004, at 10:20 AM, anatoli_krassavine wrote:
> > 
> > > Hello Bharath,
> > >
> > > This is /grammar/ORACLE-7-SQL from ANTLR website virtually 
> verbatium.
> > >
> > > I cannot see anything in the grammar which would have caused 
such
> > > behaviour. It also appears specific to
> > >
> > > select ....... where ... in (select .....)
> > >
> > > statements
> > >
> > > I honestly tried to find what is wrong, but cannot see it.
> > >
> > > Cheers,
> > >  Toly
> > >
> > > --- In antlr-interest at yahoogroups.com, "Bharath Sundararaman"
> > > <bharath at s...> wrote:
> > >>
> > >> Hi,
> > >>
> > >> Where is the grammar for the statement? If you had a grammar 
rule
> > > which
> > >> expects a CLOSE_PAREN right after the IDENTIFIER (abc), it 
should
> > > give the
> > >> error. It would be nice if you copy-paste the rule for this SQL
> > > statement.
> > >>
> > >> Bharath.
> > >> ~Give me the tool and I shall move the earth~
> > >>
> > >> -----Original Message-----
> > >> From: anatoli_krassavine [mailto:toly at i...]
> > >> Sent: Tuesday, June 01, 2004 10:47 AM
> > >> To: antlr-interest at yahoogroups.com
> > >> Subject: [antlr-interest] Very weird SQL parser failures - 
look-
> > > ahead?
> > >>
> > >>
> > >>   Hello,
> > >>
> > >> I tried to use Oracle 7 SQL grammar in anger. It works great 
for
> > > very
> > >> complex SQL statements, but I suddenly stumbled on an 
unexpected
> > >> issues with what I consider to be a very simple SQL. Could 
> someone
> > >> please advice.
> > >>
> > >> The following statement fails with
> > >> Exception in thread "main" line 10:2: expecting CLOSE_PAREN,
> > >> found 'null'
> > >>
> > >> ------------
> > >> SELECT *
> > >> FROM
> > >>     c.abcd
> > >> WHERE
> > >>     cr.no IN
> > >> (
> > >> SELECT cr.no FROM c.cr cr
> > >> union
> > >> SELECT a FROM abc where a=b
> > >> )
> > >> ------------
> > >> On the other hand this works fine (note that "where a=b" was
> > > removed)
> > >>
> > >> SELECT *
> > >> FROM
> > >>     c.abcd
> > >> WHERE
> > >>     cr.no IN
> > >> (
> > >> SELECT cr.no FROM c.cr cr
> > >> union
> > >> SELECT a FROM abc
> > >> )
> > >> ------------
> > >>
> > >> AFAICC, it is not a problem with syntax, but a problem with a 
> look-
> > > ahead,
> > >> but I do not understand how to fix this.
> > >>
> > >> Any help or pointing in the right direction will be greatly
> > >> appreciated.
> > >>
> > >> Details:
> > >> Oracle 7 SQL grammar from  /grammar/ORACLE-7-SQL
> > >> No changes, conventional lexer/parser generated
> > >>
> > >>
> > >>
> > >>
> > >>
> > >> Yahoo! Groups Links
> > >
> > >
> > >
> > >
> > > Yahoo! Groups Links
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >



 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/antlr-interest/

<*> To unsubscribe from this group, send an email to:
     antlr-interest-unsubscribe at yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
     http://docs.yahoo.com/info/terms/