|
| 1 | +/* |
| 2 | + * Copyright 2021 the original author or authors. |
| 3 | + * |
| 4 | + * Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | + * you may not use this file except in compliance with the License. |
| 6 | + * You may obtain a copy of the License at |
| 7 | + * |
| 8 | + * https://www.apache.org/licenses/LICENSE-2.0 |
| 9 | + * |
| 10 | + * Unless required by applicable law or agreed to in writing, software |
| 11 | + * distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | + * See the License for the specific language governing permissions and |
| 14 | + * limitations under the License. |
| 15 | + */ |
| 16 | + |
| 17 | +package io.r2dbc.postgresql.util.sql; |
| 18 | + |
| 19 | +import java.util.ArrayList; |
| 20 | +import java.util.Arrays; |
| 21 | +import java.util.List; |
| 22 | + |
| 23 | +import static java.lang.Character.isDigit; |
| 24 | +import static java.lang.Character.isWhitespace; |
| 25 | +import static java.lang.Character.toLowerCase; |
| 26 | + |
| 27 | +public class BasicPostgresqlSqlLexer { |
| 28 | + |
| 29 | + private static final char[] SPECIAL_AND_OPERATOR_CHARS = { |
| 30 | + '+', '-', '*', '/', '<', '>', '=', '~', '!', '@', '#', '%', '^', '&', '|', '`', '?', |
| 31 | + '(', ')', '[', ']', ',', ';', ':', '*', '.', '\'', '"' |
| 32 | + }; |
| 33 | + |
| 34 | + static { |
| 35 | + Arrays.sort(SPECIAL_AND_OPERATOR_CHARS); |
| 36 | + } |
| 37 | + |
| 38 | + public static TokenizedSql tokenize(String sql) { |
| 39 | + List<Token> tokens = new ArrayList<>(); |
| 40 | + List<TokenizedStatement> statements = new ArrayList<>(); |
| 41 | + |
| 42 | + int statementStartIndex = 0; |
| 43 | + int i = 0; |
| 44 | + while (i < sql.length()) { |
| 45 | + char c = sql.charAt(i); |
| 46 | + Token token = null; |
| 47 | + |
| 48 | + if (isWhitespace(c)) { |
| 49 | + i++; |
| 50 | + continue; |
| 51 | + } |
| 52 | + switch (c) { |
| 53 | + case '\'': // "Standard" string constant |
| 54 | + token = getStandardQuoteToken(sql, i); |
| 55 | + break; |
| 56 | + case '\"': // Quoted Identifier |
| 57 | + token = getQuotedIdentifierToken(sql, i); |
| 58 | + break; |
| 59 | + case '-': // Possible start of double-dash comment |
| 60 | + if ((i + 1) < sql.length() && sql.charAt(i + 1) == '-') { |
| 61 | + token = getCommentToLineEndToken(sql, i); |
| 62 | + } |
| 63 | + break; |
| 64 | + case '/': // Possible start of c-style comment |
| 65 | + if ((i + 1) < sql.length() && sql.charAt(i + 1) == '*') { |
| 66 | + token = getBlockCommentToken(sql, i); |
| 67 | + } |
| 68 | + break; |
| 69 | + case '$': // Dollar-quoted constant or parameter |
| 70 | + token = getParameterOrDollarQuoteToken(sql, i); |
| 71 | + break; |
| 72 | + case ';': |
| 73 | + token = new Token(TokenType.STATEMENT_END, ";"); |
| 74 | + break; |
| 75 | + default: |
| 76 | + break; |
| 77 | + } |
| 78 | + if (token == null) { |
| 79 | + if (isSpecialOrOperatorChar(c)) { |
| 80 | + token = new Token(TokenType.SPECIAL_OR_OPERATOR, Character.toString(c));//getSpecialOrOperatorToken(sql, i); |
| 81 | + } else { |
| 82 | + token = getDefaultToken(sql, i); |
| 83 | + } |
| 84 | + } |
| 85 | + |
| 86 | + i += token.getValue().length(); |
| 87 | + |
| 88 | + if (token.getType() == TokenType.STATEMENT_END) { |
| 89 | + |
| 90 | + tokens.add(token); |
| 91 | + statements.add(new TokenizedStatement(sql.substring(statementStartIndex, i), tokens)); |
| 92 | + |
| 93 | + tokens = new ArrayList<>(); |
| 94 | + statementStartIndex = i + 1; |
| 95 | + } else { |
| 96 | + tokens.add(token); |
| 97 | + } |
| 98 | + } |
| 99 | + // If tokens is not empty, implicit statement end |
| 100 | + if (!tokens.isEmpty()) { |
| 101 | + statements.add(new TokenizedStatement(sql.substring(statementStartIndex), tokens)); |
| 102 | + } |
| 103 | + |
| 104 | + return new TokenizedSql(sql, statements); |
| 105 | + } |
| 106 | + |
| 107 | + private static Token getDefaultToken(String sql, int beginIndex) { |
| 108 | + for (int i = beginIndex + 1; i < sql.length(); i++) { |
| 109 | + char c = sql.charAt(i); |
| 110 | + if (Character.isWhitespace(c) || isSpecialOrOperatorChar(c)) { |
| 111 | + return new Token(TokenType.DEFAULT, sql.substring(beginIndex, i)); |
| 112 | + } |
| 113 | + } |
| 114 | + return new Token(TokenType.DEFAULT, sql.substring(beginIndex)); |
| 115 | + } |
| 116 | + |
| 117 | + private static boolean isSpecialOrOperatorChar(char c) { |
| 118 | + return Arrays.binarySearch(SPECIAL_AND_OPERATOR_CHARS, c) >= 0; |
| 119 | + } |
| 120 | + |
| 121 | + private static Token getBlockCommentToken(String sql, int beginIndex) { |
| 122 | + int depth = 1; |
| 123 | + for (int i = beginIndex + 2; i < (sql.length() - 1); i++) { |
| 124 | + String biGraph = sql.substring(i, i + 2); |
| 125 | + |
| 126 | + if (biGraph.equals("/*")) { |
| 127 | + depth++; |
| 128 | + i++; |
| 129 | + } else if (biGraph.equals("*/")) { |
| 130 | + depth--; |
| 131 | + i++; |
| 132 | + } |
| 133 | + if (depth == 0) { |
| 134 | + return new Token(TokenType.COMMENT, sql.substring(beginIndex, i + 1)); |
| 135 | + } |
| 136 | + } |
| 137 | + throw new IllegalArgumentException("Sql cannot be parsed: unclosed block comment (comment opened at index " + beginIndex + ") in statement: " + sql); |
| 138 | + } |
| 139 | + |
| 140 | + private static Token getCommentToLineEndToken(String sql, int beginIndex) { |
| 141 | + int lineEnding = sql.indexOf('\n', beginIndex); |
| 142 | + if (lineEnding == -1) { |
| 143 | + return new Token(TokenType.COMMENT, sql.substring(beginIndex)); |
| 144 | + } else { |
| 145 | + return new Token(TokenType.COMMENT, sql.substring(beginIndex, lineEnding)); |
| 146 | + } |
| 147 | + } |
| 148 | + |
| 149 | + private static Token getDollarQuoteToken(String sql, String tag, int beginIndex) { |
| 150 | + int nextQuote = sql.indexOf(tag, beginIndex + tag.length()); |
| 151 | + if (nextQuote == -1) { |
| 152 | + throw new IllegalArgumentException("Sql cannot be parsed: unclosed quote (quote opened at index " + beginIndex + ") in statement: " + sql); |
| 153 | + } else { |
| 154 | + return new Token(TokenType.STRING_CONSTANT, sql.substring(beginIndex, nextQuote + tag.length())); |
| 155 | + } |
| 156 | + } |
| 157 | + |
| 158 | + private static Token getParameterToken(String sql, int beginIndex) { |
| 159 | + for (int i = beginIndex + 1; i < sql.length(); i++) { |
| 160 | + char c = sql.charAt(i); |
| 161 | + if (isWhitespace(c) || isSpecialOrOperatorChar(c)) { |
| 162 | + return new Token(TokenType.PARAMETER, sql.substring(beginIndex, i)); |
| 163 | + } |
| 164 | + if (!isAsciiDigit(c)) { |
| 165 | + throw new IllegalArgumentException("Sql cannot be parsed: illegal character in parameter or dollar-quote tag: " + c); |
| 166 | + } |
| 167 | + } |
| 168 | + return new Token(TokenType.PARAMETER, sql.substring(beginIndex)); |
| 169 | + } |
| 170 | + |
| 171 | + private static Token getParameterOrDollarQuoteToken(String sql, int beginIndex) { |
| 172 | + char firstChar = sql.charAt(beginIndex + 1); |
| 173 | + if (firstChar == '$') { |
| 174 | + return getDollarQuoteToken(sql, "$$", beginIndex); |
| 175 | + } else if (isAsciiDigit(firstChar)) { |
| 176 | + return getParameterToken(sql, beginIndex); |
| 177 | + } else { |
| 178 | + for (int i = beginIndex + 1; i < sql.length(); i++) { |
| 179 | + char c = sql.charAt(i); |
| 180 | + if (c == '$') { |
| 181 | + return getDollarQuoteToken(sql, sql.substring(beginIndex, i + 1), beginIndex); |
| 182 | + } |
| 183 | + if (!(isAsciiLetter(c) || c == '_' || isAsciiDigit(c))) { |
| 184 | + throw new IllegalArgumentException("Sql cannot be parsed: illegal character in dollar-quote tag (quote opened at index " + beginIndex + ") in statement: " + sql); |
| 185 | + } |
| 186 | + } |
| 187 | + throw new IllegalArgumentException("Sql cannot be parsed: unclosed dollar-quote tag(quote opened at index " + beginIndex + ") in statement: " + sql); |
| 188 | + } |
| 189 | + } |
| 190 | + |
| 191 | + private static Token getStandardQuoteToken(String sql, int beginIndex) { |
| 192 | + int nextQuote = sql.indexOf('\'', beginIndex + 1); |
| 193 | + if (nextQuote == -1) { |
| 194 | + throw new IllegalArgumentException("Sql cannot be parsed: unclosed quote (quote opened at index " + beginIndex + ") in statement: " + sql); |
| 195 | + } else { |
| 196 | + return new Token(TokenType.STRING_CONSTANT, sql.substring(beginIndex, nextQuote + 1)); |
| 197 | + } |
| 198 | + } |
| 199 | + |
| 200 | + private static Token getQuotedIdentifierToken(String sql, int beginIndex) { |
| 201 | + int nextQuote = sql.indexOf('\"', beginIndex + 1); |
| 202 | + if (nextQuote == -1) { |
| 203 | + throw new IllegalArgumentException("Sql cannot be parsed: unclosed quoted identifier (identifier opened at index " + beginIndex + ") in statement: " + sql); |
| 204 | + } else { |
| 205 | + return new Token(TokenType.QUOTED_IDENTIFIER, sql.substring(beginIndex, nextQuote + 1)); |
| 206 | + } |
| 207 | + } |
| 208 | + |
| 209 | + private static boolean isAsciiLetter(char c){ |
| 210 | + char lower = Character.toLowerCase(c); |
| 211 | + return lower >= 'a' && lower <= 'z'; |
| 212 | + } |
| 213 | + |
| 214 | + private static boolean isAsciiDigit(char c){ |
| 215 | + return c >= '0' && c <= '9'; |
| 216 | + } |
| 217 | +} |
0 commit comments