Strip quotes out from column and table names

src/column-parser.ts

@@ -1,4 +1,5 @@
 import { ColumnReference, Dialect, Token } from './defines';
+import { maybeIdentifier, maybeStripQuotes } from './utils';
 
 // States for skipping MSSQL's TOP clause: SELECT TOP n [PERCENT] [WITH TIES]
 // The tokenizer emits digits as individual single-character 'unknown' tokens,
@@ -255,7 +256,7 @@ export class ColumnParser {
           prevNonWhitespaceToken?.value !== '.' &&
           prevNonWhitespaceToken?.value !== ',' &&
           prevToken?.type === 'whitespace' &&
-          this.maybeIdent(token)
+          maybeIdentifier(token, this.dialect)
         ) {
           if (!this.alias) {
             this.alias = token.value;
@@ -300,22 +301,22 @@ export class ColumnParser {
     if (this.parts.length === 1) {
       const name = this.parts[0];
       col = {
-        name,
+        name: maybeStripQuotes(name, this.dialect),
         isWildcard: name === '*',
       };
     } else if (this.parts.length === 2) {
       const [table, name] = this.parts;
       col = {
-        name,
-        table,
+        name: maybeStripQuotes(name, this.dialect),
+        table: maybeStripQuotes(table, this.dialect),
         isWildcard: name === '*',
       };
     } else if (this.parts.length === 3) {
       const [schema, table, name] = this.parts;
       col = {
-        name,
-        table,
-        schema,
+        name: maybeStripQuotes(name, this.dialect),
+        table: maybeStripQuotes(table, this.dialect),
+        schema: maybeStripQuotes(schema, this.dialect),
         isWildcard: name === '*',
       };
     } else {
@@ -327,7 +328,7 @@ export class ColumnParser {
     }
 
     if (!!this.alias && !!col) {
-      col.alias = this.alias;
+      col.alias = maybeStripQuotes(this.alias, this.dialect);
     }
 
     return col;
@@ -346,10 +347,4 @@ export class ColumnParser {
       col.alias ?? 'none'
     }`;
   }
-
-  private maybeIdent(token: Token): boolean {
-    const ch = token.value[0];
-    const startChars = this.dialect === 'mssql' ? ['"', '['] : ['"', '`'];
-    return token.type !== 'string' && (startChars.includes(ch) || /[a-zA-Z_]/.test(ch));
-  }
 }

src/parser.ts

@@ -856,7 +856,7 @@ function stateMachineStatementParser(
   let openBlocks = 0;
 
   const columnParser = new ColumnParser(dialect);
-  const tableParser = new TableParser();
+  const tableParser = new TableParser(dialect);
 
   /* eslint arrow-body-style: 0, no-extra-parens: 0 */
   const isValidToken = (step: Step, token: Token) => {

src/table-parser.ts

@@ -1,4 +1,5 @@
-import { TableReference, Token } from './defines';
+import { Dialect, TableReference, Token } from './defines';
+import { maybeStripQuotes } from './utils';
 
 export class TableParser {
   private parts: string[] = [];
@@ -9,6 +10,8 @@ export class TableParser {
   private maybeCommaSep = false;
   private parensDepth = 0;
 
+  constructor(private dialect: Dialect) {}
+
   // keywords that come directly before a table name.
   // v1 - keeping it very simple.
   private PRE_TABLE_KEYWORDS = new Set<string>(['FROM', 'JOIN', 'INTO']);
@@ -140,20 +143,20 @@ export class TableParser {
     if (this.parts.length === 1) {
       const name = this.parts[0];
       table = {
-        name,
+        name: maybeStripQuotes(name, this.dialect),
       };
     } else if (this.parts.length === 2) {
       const [schema, name] = this.parts;
       table = {
-        name,
-        schema,
+        name: maybeStripQuotes(name, this.dialect),
+        schema: maybeStripQuotes(schema, this.dialect),
       };
     } else if (this.parts.length === 3) {
       const [database, schema, name] = this.parts;
       table = {
-        name,
-        schema,
-        database,
+        name: maybeStripQuotes(name, this.dialect),
+        schema: maybeStripQuotes(schema, this.dialect),
+        database: maybeStripQuotes(database, this.dialect),
       };
     } else {
       const fullName = this.parts.join('.');
@@ -163,7 +166,7 @@ export class TableParser {
     }
 
     if (!!this.alias && !!table) {
-      table.alias = this.alias;
+      table.alias = maybeStripQuotes(this.alias, this.dialect);
     }
 
     return table;

src/tokenizer.ts

@@ -3,6 +3,7 @@
  */
 
 import type { Token, State, Dialect, ParamTypes } from './defines';
+import { getStartQuotes } from './utils';
 
 type Char = string | null;
 
@@ -117,7 +118,7 @@ export function scanToken(
   }
 
   if (isQuotedIdentifier(ch, dialect) && ch !== null) {
-    return scanQuotedIdentifier(state, ENDTOKENS[ch]);
+    return scanQuotedIdentifier(state, ENDTOKENS[ch], dialect);
   }
 
   if (isLetter(ch)) {
@@ -385,11 +386,24 @@ function scanCommentBlock(state: State): Token {
   };
 }
 
-function scanQuotedIdentifier(state: State, endToken: Char): Token {
-  let nextChar: Char;
-  do {
+function scanQuotedIdentifier(state: State, endToken: Char, dialect: Dialect): Token {
+  let nextChar: Char = peek(state);
+  while (nextChar !== null) {
     nextChar = read(state);
-  } while (endToken !== nextChar && nextChar !== null);
+    if (nextChar === null) break;
+
+    if (nextChar === endToken && peek(state) === endToken) {
+      read(state);
+      continue;
+    }
+
+    if (dialect === 'bigquery' && nextChar === '\\' && peek(state) === endToken) {
+      read(state);
+      continue;
+    }
+
+    if (nextChar === endToken) break;
+  }
 
   if (nextChar !== null && endToken !== nextChar) {
     unread(state);
@@ -520,7 +534,7 @@ function isDollarQuotedString(state: State): boolean {
 }
 
 function isQuotedIdentifier(ch: Char, dialect: Dialect): boolean {
-  const startQuoteChars: Char[] = dialect === 'mssql' ? ['"', '['] : ['"', '`'];
+  const startQuoteChars: Char[] = getStartQuotes(dialect);
   return startQuoteChars.includes(ch);
 }
 

src/utils.ts

@@ -0,0 +1,48 @@
+import { Dialect, Token } from './defines';
+
+export function getStartQuotes(dialect: Dialect): string[] {
+  if (dialect === 'mssql') {
+    return ['"', '['];
+  } else {
+    return ['"', '`'];
+  }
+}
+
+function endQuoteFor(char: string): string {
+  if (char === '[') {
+    return ']';
+  }
+  return char;
+}
+
+export function maybeIdentifier(token: Token, dialect: Dialect): boolean {
+  const ch = token.value[0];
+  const startChars = getStartQuotes(dialect);
+  return token.type !== 'string' && (startChars.includes(ch) || /[a-zA-Z_]/.test(ch));
+}
+
+export function maybeStripQuotes(value: string, dialect: Dialect): string {
+  if (value.length < 2) {
+    return value;
+  }
+
+  const start = value[0];
+  const end = value[value.length - 1];
+
+  if (!getStartQuotes(dialect).includes(start)) {
+    return value;
+  }
+
+  const expectedEnd = endQuoteFor(start);
+  if (end !== expectedEnd) {
+    return value;
+  }
+
+  const inner = value.slice(1, -1);
+
+  if (dialect === 'bigquery' && start === '`') {
+    return inner.replace(/\\`/g, '`');
+  }
+
+  return inner.split(expectedEnd + expectedEnd).join(expectedEnd);
+}