我想通过 boost qi 实现 SQL WHERE 子句解析器的一些规则。如果在https://jakewheat.github.io/sql-overview/sql-2016-foundation-grammar.html#_8_4_in_predicate上找到它的 BNF 。我通过一些特殊的操作数实现了特定的谓词规则。但是当我想将它与 OR/AND/NOT 逻辑运算符结合使用时，它不起作用。以下源代码测试用例和语法。
#include <iostream>
#include <fstream>
#include <vector>
#include <string>
#include <variant>
#include <functional>
#include <tuple>
#include <boost/spirit/include/qi.hpp>
#include <boost/spirit/include/qi_grammar.hpp>
#include <boost/spirit/repository/include/qi_distinct.hpp>
const std::vector<std::tuple<bool /* parsing result */, std::string /* expr */>> testcases_{
{true, "(a between (12,13))",},
{true, "(a between [12,13))",},
{true, "(a between [12,13])",},
{true, "(a between (12,13])",},
{true, "(a not between (12,13))",},
{true, "(a not between [12,13))",},
{true, "(a not between [12,13])",},
{true, "(a not between (12,13])",},
{true, "(a between ('12',13))",},
{true, "(a between ['12',13))",},
{true, "(a between ['12',13])",},
{true, "(a between ('12',13])",},
{true, "(a not between ('12',13))",},
{true, "(a not between ['12',13))",},
{true, "(a not between ['12',13])",},
{true, "(a not between ('12',13])",},
{true, "(a between ('12','13'))",},
{true, "(a between ['12','13'))",},
{true, "(a between ['12','13'])",},
{true, "(a between ('12','13'])",},
{true, "(a not between ('12','13'))",},
{true, "(a not between ['12','13'))",},
{true, "(a not between ['12','13'])",},
{true, "(a not between ('12','13'])",},
{true, "(a in (12,13,3433))",},
{true, "(a in (12,13,3433))",},
{true, "a in (12,13,3433)",},
{true, "a in ('12','13','3433')",},
{true, "(a in ('12','13','3433'))",},
{true, "(a NOT in (12,13,3433))",},
{true, "a not in (12,13,3433)",},
{true, "a NOT in ('12','13','3433')",},
{true, "(a not in ('12','13','3433'))",},
{true, "a gt 'b'",},
{true, "a > 'b'",},
{true, "(a gt 'b')",},
{true, "(a > 'b')",},
{true, "a >= 'b'",},
{true, "a => 'b'",},
{true, "a gte 'b'",},
{true, "(a gte 'b')",},
{true, "(a >= 'b')",},
{true, "(a => 'b')",},
{true, "a lt 'b'",},
{true, "a < 'b'",},
{true, "a <= 'b'",},
{true, "a =< 'b'",},
{true, "a lte 'b'",},
{true, "(a lte 'b')",},
{true, "(a <= 'b')",},
{true, "(a =< 'b')",},
{true, "a eq 122",},
{true, "a == 150",},
{true, "a = 0",},
{true, "a eq '122'",},
{true, "a == '150'",},
{true, "a = '0'",},
{true, "a neq 122",},
{true, "a <> 150",},
{true, "a != 0",},
{true, "a neq '122'",},
{true, "a <> '150'",},
{true, "a != '0'",},
{true, "a like 'b'",},
{true, "a not like 'b'",},
{true, "(a like 'b')",},
{true, "(a not like 'b')",},
{true, "a like '*b'",},
{true, "a not like 'b#32'",},
{true, "(a like 'bdfwdf')",},
{true, "(a not like 'b888..///')",},
}
;
namespace sql {
namespace qi = boost::spirit::qi;
using namespace qi;
using boost::spirit::repository::distinct;
template <typename Iterator, typename Signature= unused_type(), typename Skipper= unused_type, typename Locals= unused_type >
struct parser_t: grammar<Iterator, Signature, Skipper, Locals>{
parser_t():
parser_t::base_type(start_)
{
//---------------------------------------------------------------------------
identifier_= qi::no_case[ qi::lexeme[ (qi::alpha >> *(qi::alnum | '_')) ] ] ;
str_val_= qi::lexeme[ qi::char_('\'') >> +( qi::char_ - qi::char_('\'') ) >> qi::char_('\'')];
num_val_ = +digit;
value_= ( num_val_ | str_val_ );
in_value_= lit("(") >> ((num_val_ % ',') | (str_val_ % ',')) >> lit(")");
//---------------------------------------------------------------------------
AND= (no_case[lit("AND")] |
lit("&&") |
distinct(lit("&"))[lit("&")]);
OR= (no_case[lit("OR")] |
lit("||") |
distinct(lit("|"))[lit("|")]);
NOT= (no_case[lit("NOT")] |
distinct(lit("="))[lit("!")]);
BETWEEN= no_case[lit("BETWEEN")];
IN= no_case[lit("IN")];
LIKE= no_case[lit("LIKE")];
IS= (no_case[lit("IS")]);
TRUE= no_case[lit("TRUE")];
FALSE= no_case[lit("FALSE")];
truth_value_= ( TRUE | FALSE);
//---------------------------------------------------------------------------
left_paren_= lit("(");
right_paren_= lit(")");
left_bracket_= lit("[");
right_bracket_= lit("]");
comma_= lit(",");
// eq, =, ==
equals_operator_=
no_case[lit("eq")] |
distinct(lit(">") | lit("<")| lit("="))[lit("=")] |
lit("==");
// neq, !=, <>
not_equals_operator_=
no_case[lit("neq")] |
lit("!=") |
lit("<>");
// lt, <
less_than_operator_=
distinct(no_case[lit("e")])[no_case[lit("lt")]] |
distinct('=')[lit("<")];
// gt, >
greater_than_operator_=
distinct(no_case[lit("e")])[no_case[lit("gt")]] |
distinct('=')[lit(">")];
// lte, <=, =<
less_than_or_equals_operator_=
no_case[lit("lte")] |
lit("<=") |
lit("=<");
//gte, >=, =>
greater_than_or_equals_operator_=
no_case[lit("gte")] |
lit(">=") |
lit("=>");
//---------------------------------------------------------------------------
between_predicate_= identifier_ >> -( NOT ) >> BETWEEN >> ( left_paren_ | left_bracket_ ) >> value_ >> comma_ >> value_ >> ( right_paren_ | right_bracket_) ;
in_predicate_= identifier_ >> -( NOT ) >> IN >> in_value_;
like_predicate_= identifier_ >> -( NOT ) >> LIKE >> str_val_;
comp_op_= equals_operator_ | not_equals_operator_ | less_than_operator_ | greater_than_operator_ | less_than_or_equals_operator_ | greater_than_or_equals_operator_;
comparison_predicate_= identifier_ >> comp_op_ >> value_;
predicate_= comparison_predicate_ | between_predicate_ | in_predicate_ | like_predicate_ ;
//---------------------------------------------------------------------------
start_ = or_.alias();
or_ = and_ >> *( OR >> and_);
and_ = not_ >> *( AND >> not_);
not_ = (NOT >> simple_) | simple_;
simple_ = (left_paren_ >> start_ >> right_paren_) | predicate_;
/*
boolean_predicand_=left_paren_ >> boolean_value_expression_ >> right_paren_;
boolean_primary_= predicate_ | boolean_predicand_;
boolean_test_= boolean_primary_ >> -( IS >> -( NOT ) >> truth_value_ );
boolean_factor_= -( NOT ) >> boolean_test_;
boolean_term_= boolean_factor_ | (boolean_term_ >> AND >> boolean_factor_);
boolean_value_expression_= boolean_term_ | ( boolean_value_expression_ >> OR >> boolean_term_);
start_= boolean_value_expression_;
*/
//start_= predicate_ | (start_ >> OR >> predicate_);
//---------------------------------------------------------------------------
}
rule<Iterator, Signature, Skipper, Locals> start_;
rule<Iterator, Signature, Skipper, Locals>
boolean_value_expression_,
boolean_term_,
boolean_factor_,
boolean_test_,
boolean_primary_,
boolean_predicand_
;
rule<Iterator, Signature, Skipper, Locals>
or_,
and_,
not_,
simple_
;
rule<Iterator, Signature, Skipper, Locals>
predicate_,
comparison_predicate_,
comp_op_,
between_predicate_,
in_predicate_,
like_predicate_
;
rule<Iterator, Signature, Skipper, Locals>
equals_operator_,
not_equals_operator_,
less_than_operator_,
greater_than_operator_,
less_than_or_equals_operator_,
greater_than_or_equals_operator_
;
rule<Iterator, Signature, Skipper, Locals>
AND, OR, IS, NOT,
BETWEEN, IN, LIKE,
TRUE, FALSE, truth_value_
;
rule<Iterator, Signature, Skipper, Locals>
left_paren_,
right_paren_,
left_bracket_,
right_bracket_,
comma_
;
rule<Iterator, Signature, Skipper, Locals>
identifier_,
str_val_,
num_val_,
value_,
in_value_
;
};
}
int main()
{
namespace qi = boost::spirit::qi;
using namespace qi;
// Single predicates
for(auto testcase: testcases_)
{
auto expected_result{std::get<0>(testcase)};
auto expr{std::get<1>(testcase)};
auto error(false);
std::cout << "---------------------------------------------------------------------" << std::endl <<
"Expr: " << expr << std::endl;
auto f(std::begin(expr)), l(std::end(expr));
sql::parser_t<decltype(f),
unused_type(),
qi::space_type,
unused_type> parser;
try
{
bool ok = qi::phrase_parse(f, l, parser, boost::spirit::qi::space);
if (!ok)
{
error= true;
std::cerr << "invalid expr\n";
}
else
std::cout << "result: " /*<< result */<< "\n";
} catch (const qi::expectation_failure<decltype(f)>& e)
{
error= true;
std::cerr << "expectation_failure at '" << std::string(e.first, e.last) << "'\n";
} catch(...)
{
error= true;
std::cerr << "Excepion!" << std::endl;
}
if (f!=l)
{
error= true;
std::cerr << "unparsed: '" << std::string(f,l) << "'\n";
}
if(expected_result != !error && std::getchar() == 'q') break;
}
// OR/AND predicates
auto terminate(false);
for(auto testcase1: testcases_)
{
auto expected_result1{std::get<0>(testcase1)};
auto expr1{std::get<1>(testcase1)};
for(auto testcase2: testcases_)
{
auto expected_result2{std::get<0>(testcase2)};
auto expr2{std::get<1>(testcase2)};
auto expr(expr1 + "OR" + expr2);
auto expected_result(expected_result1 && expected_result2);
auto error(false);
std::cout << "---------------------------------------------------------------------" << std::endl <<
"Expr: " << expr << std::endl;
auto f(std::begin(expr)), l(std::end(expr));
sql::parser_t<decltype(f),
unused_type(),
qi::space_type,
unused_type> parser;
try
{
bool ok = qi::phrase_parse(f, l, parser, boost::spirit::qi::space);
if (!ok)
{
error= true;
std::cerr << "invalid expr\n";
}
else
std::cout << "result: " /*<< result */<< "\n";
} catch (const qi::expectation_failure<decltype(f)>& e)
{
error= true;
std::cerr << "expectation_failure at '" << std::string(e.first, e.last) << "'\n";
} catch(...)
{
error= true;
std::cerr << "Excepion!" << std::endl;
}
if (f!=l)
{
error= true;
std::cerr << "unparsed: '" << std::string(f,l) << "'\n";
}
if(expected_result != !error && std::getchar() == 'q')
{
terminate=true;
break;
}
}
if(terminate)break;
}
return 0;
}