[antlr-interest] SQL grammar

[wirving at beats.hu]

Hi,

>From "sequence" I guess it's PL/SQL. If it is, then a PL/SQL grammar is a
good starting point: http://antlr.org/grammar/1209225566284/PLSQL3.g

Although I haven't tried this yet.


   Zoli


> Hi,
>
> Could anybody help me with writing a grammar that does the following:
>
> Parsing values that are present after VALUES key world in the INSERT INTO
> Table(column1,...) VALUES(value1,...);
>
> A value1 might have 3 different form:
>
>    1.    DEFAULT  (any string, e.g. USER, CURRENT SQLID)
>    2.     NULL (e.g. INSERT INTO Table(column1,column2) VALUES(NULL,NULL);
>    3.     expressions
>
>     Operators (|| , /, + , -, *) might be used with expressions (they are
> optional)
>
> a.       scalar
>
> o    date: SYSDATE, CURRENT DATE
>
> o    string: 'a ' b'
>
> o    number: 6.2
>
> o    column name: column1
>
> b.       Labeled durations
>
> o    HIREDATE + 2 MONTHS + 14 DAYS
>
> c.        sequence-reference
>
> o    sequence-name.nextval
>
> o    NEXT VALUE FOR sequence-name
>
> d.       CASE expression
>
> o    CASE searched-when-clause ELSE result-expression END
>
> o    WHEN search-condition THEN result-expression
>
> o    expression WHEN expression THEN result-expression
>
> e.        CAST expression
>
> o    CAST ( expression AS datatype)
>
> f.        function-invocation
>
> A function input parameter might be any argument form sections a-d or
> another function
>
> o    CONCAT('1', '1', 'd')
>
> o    CONCAT(col1, 'f',1)
>
> o    package.name.function(5)
>
>
>
>
>
> I have combined the complex example with all the cases included:
>
>
>
> INSERT INTO
> TABLE(c1,c2,c3,c4,c5,c6,c7,c8,c9,c10,c11,c12,c13,c14,c15,c16,c17,c18,c19)
> VALUES ('Don''t Pay' , 'Used to haul fuel , water'  , '' , ,, NULL , 1 ,
> '1'||chr(10)||'COUNTRY=GRB' , CURRENT TIMESTAMP , CURRENT DATE +1 ,
> HIREDATE
> + 2 MONTHS + 14 DAYS ,  CONCAT(c1 , 'x' || CURENT DATE  , 2 + 2 , f( 7 ,
> 'a
> '' b',p.n.f.g('i' ) ) ) , (2 + 8 * 16) / 2 - 3 , sequence-name.nextval ,
> NEXT VALUE FOR sequence-name , case when b = '*' then 'star' when b = '+'
> then 'plus' when b = '-' then 'minus' else '?''?' end ,  ( CASE WHEN
> FIELDNAME IS NULL THEN DEFAULTVALUE ELSE FIELDNAME END ) , CAST (
�1� AS
> NUMBER ), SELECT 'i''d='||id AS ID FROM (SELECT T1.id from T1 UNION ALL
> Select t2.id from T2 t2));
>
>
>
> I expect to get:
>
>
>
> c1 -> 'Don''t Pay'
>
> c2 -> 'Used to haul fuel , water'
>
> c3 -> ''
>
> c4 ->
>
> c5 ->
>
> c6 -> NULL
>
> c7 ->  1
>
> c8 -> '1'||chr(10)||'COUNTRY=GRB'
>
> c9 -> CURRENT TIMESTAMP
>
> c10 -> CURRENT DATE +1
>
> c11 -> HIREDATE + 2 MONTHS + 14 DAYS
>
> c12 -> CONCAT(c1 , 'x' || CURENT DATE  , 2 + 2 , f( 7 , 'a '
> b',p.n.f.g('i'
> ) ) )
>
> c13 -> (2 + 8 * 16) / 2 - 3
>
> c14 -> sequence-name.nextval
>
> c15 -> NEXT VALUE FOR sequence-name
>
> c16 -> case when b = '*' then 'star' when b = '+' then 'plus' when b = '-'
> then 'minus' else '?''?' end
>
> c17 -> ( CASE WHEN FIELDNAME IS NULL THEN DEFAULTVALUE ELSE FIELDNAME END
> )
>
> c18 -> CAST ( �1� AS NUMBER )
>
> c19 -> SELECT 'i''d='||id AS ID FROM (SELECT T1.id from T1 UNION ALL
> Select
> t2.id from T2 t2)
>
>
>
> Mind that c4 is white space and c5 has no value at all.
>
>
> Thanks in advance
>
> Marcin
>
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
> Unsubscribe:
> http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>