Grammar for the SQL dialect support in ADSI's OLE DB provider. 

The grammar we used is a subset derived from the SQL ODBC miniumum grammar.

2.1 The Terminal Tokens
TERMINAL SYMBOLS (TOKENS) to be expected by the parser. 

TOKEN_ERROR              
TOKEN_STAR               
TOKEN_LPARAN             
TOKEN_RPARAN             
TOKEN_INTEGER_LITERAL    
TOKEN_REAL_LITERAL       
TOKEN_STRING_LITERAL     
TOKEN_USER_DEFINED_NAME  
TOKEN_COMMA              
TOKEN_SELECT             
TOKEN_ALL                
TOKEN_FROM               
TOKEN_WHERE              
TOKEN_BOOLEAN_LITERAL 
TOKEN_AND                
TOKEN_OR                 
TOKEN_NOT    
TOKEN_END           
TOKEN_LT
TOKEN_GT
TOKEN_GE
TOKEN_LE
TOKEN_EQ
TOKEN_NE

2.2 The Grammer
statement ::= select-statement

select-statement ::= TOKEN_SELECT [TOKEN_ALL] select-list TOKEN_FROM table-identifier     [TOKEN_WHERE search-condition] 

select-list ::= TOKEN_STAR | select-sublist [TOKEN_COMMA select-sublist]... 

select-sublist ::= column-identifier

column-identifier ::= TOKEN_USER_DEFINED_NAME 

table-identifier ::= TOKEN_STRING_LITERAL 

search-condition ::= boolean-term [TOKEN_OR search-condition]

boolean-term ::= boolean-factor [TOKEN_AND boolean-term]

boolean-factor ::= [TOKEN_NOT] boolean-primary

boolean-primary ::= comparison-predicate | TOKEN_LPARAN search-condition TOKEN_RPARAN

comparison-predicate ::= column-identifier TOKEN_COMPARISON_OPERATOR literal

literal ::= TOKEN_STRING_LITERAL | numeric-literal  | TOKEN_BOOLEAN_LITERAL 

numeric-literal ::= TOKEN_INTEGER_LITERAL | TOKEN_REAL_LITERAL



For example, the tokens generated by the lexer for the following SQL statement

SELECT CN, objectClass FROM 'LDAP://ntdsdc1/c=US/O=Microsoft/OU=NTDS' WHERE
ObjectClass = User OR AGE > 25