Re: [Hsqldb-developers] Parser / extended SQL syntax

[Johannes v. S. <jvs...@we...>]

Hi again,

fredt answered to my post on hsgldb-users so i'll follow it up here:

> In short, the parser is hand written. You can extend it if you wish to
> accept any extensions to the syntax. 
So that means, there is no parser generator of any sort which transforms 
some grammar file into corresponding java code?

> Beyond that, you need to implement the
> extra processing to return the results of fuzzy and ordinal searches.
Ok I'm aware of this. 

>HSQLDB is relatively easy to adapt but this task is not among the simplest
>ones.
I see. I think I'll look into the source the next days and see how far I get there.
Can you give maybe a hint about which classes would be affected/need to be modified?

Then I'll come back with my questions. :)

cu, Johannes

------
> 
> 
> Hello everyone,
> 
> as I already posted on hsqldb-users (sorry wasn't aware of hsqldb-developers):
> 
> 
> My question was:  Which parser hsqldb is using?
> Is it javacc etc., or a parser you have developed yourself?
> 
> Whould it be possible to change the way hsqldb parses the sql, as for 
> implementing an enhanced version of sql, with additional 
> keywords/expressions etc?
> 
> The background:
> I'm studying applied computer science in germany, and for my diploma 
> thesis, I'm enhancing an application developed at our university 
> (university of augsburg, http://www.uni-augsburg.de/).
> This application is called Preference SQL, it enhances standard SQL 
> with the notion of preferences.
> One can for example state a query like
>  SELECT * FROM usedcars
>    WHERE make = 'Ford'
>    PREFERRING price AROUND 20000
>           AND HIGHEST(power)
>           PRIOR TO color = 'red'
> for which I guess the meaning is quite obvious.
> 
> The difference to standard SQL is that the conditions in the PREFFERING 
> clause don't have to be matched exactely, but rather are treated as soft 
> selections. So if there's no exact match, the best matching entries are 
> returned. (I can give you more background about this if someone should 
> be interested  :)  ).
> 
> Until now, the program is implemented in the way that a JDBC driver 
> takes the PreferenceSQL statement, connects the the underlying 
> (standard) sql database (oracle, mysql etc), fetches all the nessesary 
> tables with hard selections (WHERE..), and executes the preference part 
> of the query on this data.
> To parse the query, we use a JavaCC generated parser.
> 
> My task would be to implement the exection of preference queries 
> directly into a java database, and I was told to use your hsqldb.
> So I would adapt the classes that do the parsing on the incoming query, 
> so that they can handle the extended syntax.
> 
> My questions would be:
> What do you think of that task and it's complexity?
> How is your parser organized/what parser do you use?
> Whould it be possible to convert our javacc parser to your parser format?
> How much of the code would I have to change to implement execution of preference queries?
> 
> many thanks for your answers,
> 
> regards, Johannes
> 
> ______________________________________________________________
> Verschicken Sie romantische, coole und witzige Bilder per SMS!
> Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> hsqldb-developers mailing list
> hsq...@li...
> https://lists.sourceforge.net/lists/listinfo/hsqldb-developers


______________________________________________________________
Verschicken Sie romantische, coole und witzige Bilder per SMS!
Jetzt bei WEB.DE FreeMail: http://f.web.de/?mc=021193