| 1 | #pragma once |
|---|---|
| 2 | |
| 3 | #include <vector> |
| 4 | #include <Parsers/Lexer.h> |
| 5 | |
| 6 | |
| 7 | namespace DB |
| 8 | { |
| 9 | |
| 10 | /** Parser operates on lazy stream of tokens. |
| 11 | * It could do lookaheads of any depth. |
| 12 | */ |
| 13 | |
| 14 | /** Used as an input for parsers. |
| 15 | * All whitespace and comment tokens are transparently skipped. |
| 16 | */ |
| 17 | class Tokens |
| 18 | { |
| 19 | private: |
| 20 | std::vector<Token> data; |
| 21 | Lexer lexer; |
| 22 | |
| 23 | public: |
| 24 | Tokens(const char * begin, const char * end, size_t max_query_size = 0) : lexer(begin, end, max_query_size) {} |
| 25 | |
| 26 | const Token & operator[] (size_t index) |
| 27 | { |
| 28 | while (true) |
| 29 | { |
| 30 | if (index < data.size()) |
| 31 | return data[index]; |
| 32 | |
| 33 | if (!data.empty() && data.back().isEnd()) |
| 34 | return data.back(); |
| 35 | |
| 36 | Token token = lexer.nextToken(); |
| 37 | |
| 38 | if (token.isSignificant()) |
| 39 | data.emplace_back(token); |
| 40 | } |
| 41 | } |
| 42 | |
| 43 | const Token & max() |
| 44 | { |
| 45 | if (data.empty()) |
| 46 | return (*this)[0]; |
| 47 | return data.back(); |
| 48 | } |
| 49 | }; |
| 50 | |
| 51 | |
| 52 | /// To represent position in a token stream. |
| 53 | class TokenIterator |
| 54 | { |
| 55 | private: |
| 56 | Tokens * tokens; |
| 57 | size_t index = 0; |
| 58 | |
| 59 | public: |
| 60 | explicit TokenIterator(Tokens & tokens_) : tokens(&tokens_) {} |
| 61 | |
| 62 | const Token & get() { return (*tokens)[index]; } |
| 63 | const Token & operator*() { return get(); } |
| 64 | const Token * operator->() { return &get(); } |
| 65 | |
| 66 | TokenIterator & operator++() { ++index; return *this; } |
| 67 | TokenIterator & operator--() { --index; return *this; } |
| 68 | |
| 69 | bool operator< (const TokenIterator & rhs) const { return index < rhs.index; } |
| 70 | bool operator<= (const TokenIterator & rhs) const { return index <= rhs.index; } |
| 71 | bool operator== (const TokenIterator & rhs) const { return index == rhs.index; } |
| 72 | bool operator!= (const TokenIterator & rhs) const { return index != rhs.index; } |
| 73 | |
| 74 | bool isValid() { return get().type < TokenType::EndOfStream; } |
| 75 | |
| 76 | /// Rightmost token we had looked. |
| 77 | const Token & max() { return tokens->max(); } |
| 78 | }; |
| 79 | |
| 80 | |
| 81 | /// Returns positions of unmatched parentheses. |
| 82 | using UnmatchedParentheses = std::vector<Token>; |
| 83 | UnmatchedParentheses checkUnmatchedParentheses(TokenIterator begin, Token * last); |
| 84 | |
| 85 | } |
| 86 |