perf: further micro optimizations to parser

Author: olavloite

google-cloud-spanner/src/main/java/com/google/cloud/spanner/Statement.java

@@ -21,6 +21,7 @@
 
 import com.google.cloud.spanner.ReadContext.QueryAnalyzeMode;
 import com.google.common.base.Preconditions;
+import com.google.common.collect.ImmutableMap;
 import com.google.spanner.v1.ExecuteSqlRequest.QueryOptions;
 import java.io.Serializable;
 import java.util.Collections;
@@ -140,7 +141,7 @@ Builder handle(Value value) {
 
   /** Creates a {@code Statement} with the given SQL text {@code sql}. */
   public static Statement of(String sql) {
-    return newBuilder(sql).build();
+    return new Statement(sql, ImmutableMap.of(), /*queryOptions=*/ null);
   }
 
   /** Creates a new statement builder with the SQL text {@code sql}. */

google-cloud-spanner/src/main/java/com/google/cloud/spanner/connection/AbstractStatementParser.java

@@ -513,7 +513,7 @@ ParsedStatement parse(Statement statement, QueryOptions defaultQueryOptions) {
     return parsedStatement.copy(statement, defaultQueryOptions);
   }
 
-  private ParsedStatement internalParse(Statement statement, QueryOptions defaultQueryOptions) {
+  ParsedStatement internalParse(Statement statement, QueryOptions defaultQueryOptions) {
     StatementHintParser statementHintParser =
         new StatementHintParser(getDialect(), statement.getSql());
     ReadQueryUpdateTransactionOption[] optionsFromHints = EMPTY_OPTIONS;
@@ -523,16 +523,21 @@ private ParsedStatement internalParse(Statement statement, QueryOptions defaultQ
           statement.toBuilder().replace(statementHintParser.getSqlWithoutClientSideHints()).build();
       optionsFromHints = convertHintsToOptions(statementHintParser.getClientSideStatementHints());
     }
+    // TODO: Qualify statements without removing comments first.
     String sql = removeCommentsAndTrim(statement.getSql());
     ClientSideStatementImpl client = parseClientSideStatement(sql);
     if (client != null) {
       return ParsedStatement.clientSideStatement(client, statement, sql);
-    } else if (isQuery(sql)) {
-      return ParsedStatement.query(statement, sql, defaultQueryOptions, optionsFromHints);
-    } else if (isUpdateStatement(sql)) {
-      return ParsedStatement.update(statement, sql, checkReturningClause(sql), optionsFromHints);
-    } else if (isDdlStatement(sql)) {
-      return ParsedStatement.ddl(statement, sql);
+    } else {
+      String sqlWithoutHints =
+          !sql.isEmpty() && sql.charAt(0) == '@' ? removeStatementHint(sql) : sql;
+      if (isQuery(sqlWithoutHints)) {
+        return ParsedStatement.query(statement, sql, defaultQueryOptions, optionsFromHints);
+      } else if (isUpdateStatement(sqlWithoutHints)) {
+        return ParsedStatement.update(statement, sql, checkReturningClause(sql), optionsFromHints);
+      } else if (isDdlStatement(sqlWithoutHints)) {
+        return ParsedStatement.ddl(statement, sql);
+      }
     }
     return ParsedStatement.unknown(statement, sql);
   }
@@ -935,7 +940,8 @@ int skipQuoted(
       appendIfNotNull(result, startQuote);
       appendIfNotNull(result, startQuote);
     }
-    while (currentIndex < sql.length()) {
+    int length = sql.length();
+    while (currentIndex < length) {
       char currentChar = sql.charAt(currentIndex);
       if (currentChar == startQuote) {
         if (supportsDollarQuotedStrings() && currentChar == DOLLAR) {
@@ -946,7 +952,7 @@ int skipQuoted(
             return currentIndex + tag.length() + 2;
           }
         } else if (supportsEscapeQuoteWithQuote()
-            && sql.length() > currentIndex + 1
+            && length > currentIndex + 1
             && sql.charAt(currentIndex + 1) == startQuote) {
           // This is an escaped quote (e.g. 'foo''bar')
           appendIfNotNull(result, currentChar);
@@ -955,7 +961,7 @@ int skipQuoted(
           continue;
         } else if (isTripleQuoted) {
           // Check if this is the end of the triple-quoted string.
-          if (sql.length() > currentIndex + 2
+          if (length > currentIndex + 2
               && sql.charAt(currentIndex + 1) == startQuote
               && sql.charAt(currentIndex + 2) == startQuote) {
             appendIfNotNull(result, currentChar);
@@ -969,7 +975,7 @@ int skipQuoted(
         }
       } else if (supportsBackslashEscape()
           && currentChar == BACKSLASH
-          && sql.length() > currentIndex + 1
+          && length > currentIndex + 1
           && sql.charAt(currentIndex + 1) == startQuote) {
         // This is an escaped quote (e.g. 'foo\'bar').
         // Note that in raw strings, the \ officially does not start an escape sequence, but the

google-cloud-spanner/src/main/java/com/google/cloud/spanner/connection/PostgreSQLStatementParser.java

@@ -125,7 +125,8 @@ String removeCommentsAndTrimInternal(String sql) {
     int multiLineCommentStartIdx = -1;
     StringBuilder res = new StringBuilder(sql.length());
     int index = 0;
-    while (index < sql.length()) {
+    int length = sql.length();
+    while (index < length) {
       char c = sql.charAt(index);
       if (isInSingleLineComment) {
         if (c == '\n') {
@@ -134,34 +135,34 @@ String removeCommentsAndTrimInternal(String sql) {
           res.append(c);
         }
       } else if (multiLineCommentLevel > 0) {
-        if (sql.length() > index + 1 && c == ASTERISK && sql.charAt(index + 1) == SLASH) {
+        if (length > index + 1 && c == ASTERISK && sql.charAt(index + 1) == SLASH) {
           multiLineCommentLevel--;
           if (multiLineCommentLevel == 0) {
-            if (!whitespaceBeforeOrAfterMultiLineComment && (sql.length() > index + 2)) {
+            if (!whitespaceBeforeOrAfterMultiLineComment && (length > index + 2)) {
               whitespaceBeforeOrAfterMultiLineComment =
                   Character.isWhitespace(sql.charAt(index + 2));
             }
             // If the multiline comment does not have any whitespace before or after it, and it is
             // neither at the start nor at the end of SQL string, append an extra space.
             if (!whitespaceBeforeOrAfterMultiLineComment
                 && (multiLineCommentStartIdx != 0)
-                && (index != sql.length() - 2)) {
+                && (index != length - 2)) {
               res.append(' ');
             }
           }
           index++;
-        } else if (sql.length() > index + 1 && c == SLASH && sql.charAt(index + 1) == ASTERISK) {
+        } else if (length > index + 1 && c == SLASH && sql.charAt(index + 1) == ASTERISK) {
           multiLineCommentLevel++;
           index++;
         }
       } else {
         // Check for -- which indicates the start of a single-line comment.
-        if (sql.length() > index + 1 && c == HYPHEN && sql.charAt(index + 1) == HYPHEN) {
+        if (length > index + 1 && c == HYPHEN && sql.charAt(index + 1) == HYPHEN) {
           // This is a single line comment.
           isInSingleLineComment = true;
           index += 2;
           continue;
-        } else if (sql.length() > index + 1 && c == SLASH && sql.charAt(index + 1) == ASTERISK) {
+        } else if (length > index + 1 && c == SLASH && sql.charAt(index + 1) == ASTERISK) {
           multiLineCommentLevel++;
           if (index >= 1) {
             whitespaceBeforeOrAfterMultiLineComment = Character.isWhitespace(sql.charAt(index - 1));

google-cloud-spanner/src/test/java/com/google/cloud/spanner/connection/StatementParserBenchmark.java

@@ -17,6 +17,9 @@
 package com.google.cloud.spanner.connection;
 
 import com.google.cloud.spanner.Dialect;
+import com.google.cloud.spanner.Statement;
+import com.google.cloud.spanner.connection.AbstractStatementParser.ParsedStatement;
+import com.google.spanner.v1.ExecuteSqlRequest.QueryOptions;
 import org.openjdk.jmh.annotations.Benchmark;
 import org.openjdk.jmh.annotations.Fork;
 import org.openjdk.jmh.annotations.Measurement;
@@ -27,28 +30,52 @@
 @Measurement(iterations = 5, time = 5)
 public class StatementParserBenchmark {
   private static final Dialect dialect = Dialect.POSTGRESQL;
-  private static final AbstractStatementParser parser =
+  private static final AbstractStatementParser PARSER =
       AbstractStatementParser.getInstance(dialect);
 
-  private static String longQueryText = generateLongQuery(100 * 1024); // 100kb
+  private static final String LONG_QUERY_TEXT =
+      generateLongStatement("SELECT * FROM foo WHERE 1", 100 * 1024); // 100kb
+
+  private static final String LONG_DML_TEXT =
+      generateLongStatement("update foo set bar=1 WHERE 1", 100 * 1024); // 100kb
 
   /** Generates a long SQL-looking string. */
-  private static String generateLongQuery(int length) {
+  private static String generateLongStatement(String prefix, int length) {
     StringBuilder sb = new StringBuilder(length + 50);
-    sb.append("SELECT * FROM foo WHERE 1");
+    sb.append(prefix);
     while (sb.length() < length) {
-      sb.append(" OR abcdefghijklmnopqrstuvwxyz");
+      sb.append(" OR abcdefghijklmnopqrstuvwxyz='abcdefghijklmnopqrstuvwxyz'");
     }
     return sb.toString();
   }
 
   @Benchmark
-  public boolean isQueryTest() {
-    return parser.isQuery("CREATE TABLE FOO (ID INT64, NAME STRING(100)) PRIMARY KEY (ID)");
+  public ParsedStatement isQueryTest() {
+    return PARSER.internalParse(
+        Statement.of("CREATE TABLE FOO (ID INT64, NAME STRING(100)) PRIMARY KEY (ID)"),
+        QueryOptions.getDefaultInstance());
+  }
+
+  @Benchmark
+  public ParsedStatement longQueryTest() {
+    return PARSER.internalParse(Statement.of(LONG_QUERY_TEXT), QueryOptions.getDefaultInstance());
   }
 
   @Benchmark
-  public boolean longQueryTest() {
-    return parser.isQuery(longQueryText);
+  public ParsedStatement longDmlTest() {
+    return PARSER.internalParse(Statement.of(LONG_DML_TEXT), QueryOptions.getDefaultInstance());
+  }
+
+  public static void main(String[] args) throws Exception {
+    for (int i = 0; i < 100000; i++) {
+      if (PARSER.internalParse(Statement.of(LONG_QUERY_TEXT), QueryOptions.getDefaultInstance())
+          == null) {
+        throw new AssertionError();
+      }
+      if (PARSER.internalParse(Statement.of(LONG_DML_TEXT), QueryOptions.getDefaultInstance())
+          == null) {
+        throw new AssertionError();
+      }
+    }
   }
 }