@chkit/core 0.1.2-beta.7 → 0.2.0-beta.11

package/dist/sql-lexer.js

@@ -0,0 +1,220 @@
+const WHITESPACE = /[ \t\n\r\f\v]/;
+const IDENTIFIER_START = /[A-Za-z_]/;
+const IDENTIFIER_PART = /[A-Za-z0-9_$]/;
+const WORD_CHAR = /[A-Za-z0-9_]/;
+const WORD = /^[A-Za-z0-9_]+$/;
+const DIGIT = /[0-9]/;
+const HEX_DIGIT = /[0-9A-Fa-f]/;
+const HEREDOC_TAG = /[A-Za-z0-9_]/;
+const EXPONENT = /^[eE][+-]?[0-9]/;
+const HEX_EXPONENT = /^[pP][+-]?[0-9]/;
+export function tokenizeSQL(sql) {
+    const tokens = [];
+    let previous;
+    let i = 0;
+    while (i < sql.length) {
+        const { kind, end, terminated } = readToken(sql, i, previous);
+        const token = { kind, text: sql.slice(i, end), start: i, end, terminated };
+        tokens.push(token);
+        if (!isTrivia(token))
+            previous = token;
+        i = end;
+    }
+    return tokens;
+}
+/** Whitespace and comments: tokens that never change what a statement means. */
+export function isTrivia(token) {
+    return token.kind === 'whitespace' || token.kind === 'line_comment' || token.kind === 'block_comment';
+}
+/** The name an identifier token denotes: bare text, or the unquoted value. */
+export function identifierName(token) {
+    if (token.kind === 'identifier')
+        return token.text;
+    if (token.kind !== 'quoted_identifier' || !token.terminated)
+        return undefined;
+    return unescapeQuoted(token.text);
+}
+/** The value of a terminated string literal (`'…'` or a `$tag$…$tag$` heredoc). */
+export function stringLiteralValue(token) {
+    if (token.kind !== 'string' || !token.terminated)
+        return undefined;
+    if (token.text.startsWith('$')) {
+        const tagLength = token.text.indexOf('$', 1) + 1;
+        return token.text.slice(tagLength, token.text.length - tagLength);
+    }
+    return unescapeQuoted(token.text);
+}
+// `previous` is the last token that is not trivia; ClickHouse uses it to tell
+// a qualifier dot from the start of a number.
+function readToken(sql, i, previous) {
+    const char = sql[i] ?? '';
+    const next = sql[i + 1] ?? '';
+    if (WHITESPACE.test(char))
+        return complete('whitespace', skipWhile(sql, i, WHITESPACE));
+    // `--x`, `//x`, `# x` and `#!x` are line comments; `#x` is not (ClickHouse rejects it).
+    if ((char === '-' && next === '-') ||
+        (char === '/' && next === '/') ||
+        (char === '#' && (next === ' ' || next === '!'))) {
+        return complete('line_comment', lineEnd(sql, i));
+    }
+    if (char === '/' && next === '*')
+        return readBlockComment(sql, i);
+    if (char === "'")
+        return { kind: 'string', ...readQuoted(sql, i, "'") };
+    if (char === '`' || char === '"')
+        return { kind: 'quoted_identifier', ...readQuoted(sql, i, char) };
+    if (char === '$') {
+        const heredocEnd = readHeredoc(sql, i);
+        if (heredocEnd !== undefined)
+            return complete('string', heredocEnd);
+        if (WORD_CHAR.test(next))
+            return complete('identifier', skipWhile(sql, i, IDENTIFIER_PART));
+    }
+    if (DIGIT.test(char))
+        return readNumberOrWord(sql, i, previous?.text === '.');
+    if (char === '.' && DIGIT.test(next) && !endsOperand(previous))
+        return complete('number', fractionEnd(sql, i));
+    if (IDENTIFIER_START.test(char))
+        return complete('identifier', skipWhile(sql, i, IDENTIFIER_PART));
+    const codePoint = sql.codePointAt(i) ?? 0;
+    return complete('punctuation', i + String.fromCodePoint(codePoint).length);
+}
+// ClickHouse reads a digit-led run as a numeric literal and, when word
+// characters continue it, as a bare word: `1m_rollup` and `2024db` are names,
+// while `1_000`, `1.5e+3` and `0x1F` are numbers. Right after a `.` only
+// integer digits form the literal, so `t.1.2` is tuple access and `db.1e5` is a
+// name. Word characters exclude `$` here: `1a$b` is `1a` followed by `$b`.
+function readNumberOrWord(sql, i, afterDot) {
+    const literalEnd = afterDot ? digitsEnd(sql, i, DIGIT) : numericLiteralEnd(sql, i);
+    const end = skipWhile(sql, literalEnd, WORD_CHAR);
+    const isName = end > literalEnd && WORD.test(sql.slice(i, end));
+    return complete(isName ? 'identifier' : 'number', end);
+}
+// An optional 0x/0b prefix, digits, an optional fraction and an optional
+// exponent. A `.` joins the literal only before a digit, so `1.x` keeps its dot.
+function numericLiteralEnd(sql, i) {
+    const prefix = sql.slice(i, i + 3);
+    const hex = /^0[xX][0-9A-Fa-f]/.test(prefix);
+    const digit = hex ? HEX_DIGIT : DIGIT;
+    let j = digitsEnd(sql, hex || /^0[bB][01]/.test(prefix) ? i + 2 : i, digit);
+    if (sql[j] === '.' && DIGIT.test(sql[j + 1] ?? ''))
+        j = digitsEnd(sql, j + 1, digit);
+    const exponent = (hex ? HEX_EXPONENT : EXPONENT).exec(sql.slice(j, j + 3));
+    return exponent ? digitsEnd(sql, j + exponent[0].length - 1, DIGIT) : j;
+}
+// One block of digits; `_` may separate two digits (`1_000`).
+function digitsEnd(sql, i, digit) {
+    let j = i;
+    while (j < sql.length) {
+        const char = sql[j] ?? '';
+        const separator = char === '_' && j > i && digit.test(sql[j + 1] ?? '');
+        if (!digit.test(char) && !separator)
+            break;
+        j += 1;
+    }
+    return j;
+}
+// `.5` or `.5e3` where a value starts.
+function fractionEnd(sql, i) {
+    const j = skipWhile(sql, i + 1, DIGIT);
+    const exponent = EXPONENT.exec(sql.slice(j, j + 3));
+    return exponent ? skipWhile(sql, j + exponent[0].length - 1, DIGIT) : j;
+}
+// After a name, a number, `)` or `]`, a `.` is a qualifier or tuple access.
+function endsOperand(token) {
+    if (token === undefined)
+        return false;
+    return (token.kind === 'identifier' ||
+        token.kind === 'quoted_identifier' ||
+        token.kind === 'number' ||
+        token.text === ')' ||
+        token.text === ']');
+}
+// ClickHouse block comments nest: `/* a /* b */ c */` is one comment.
+function readBlockComment(sql, i) {
+    let depth = 0;
+    let j = i;
+    while (j < sql.length) {
+        if (sql.startsWith('/*', j)) {
+            depth += 1;
+            j += 2;
+        }
+        else if (sql.startsWith('*/', j)) {
+            depth -= 1;
+            j += 2;
+            if (depth === 0)
+                return complete('block_comment', j);
+        }
+        else {
+            j += 1;
+        }
+    }
+    return { kind: 'block_comment', end: sql.length, terminated: false };
+}
+// A doubled delimiter or a backslash escapes the delimiter inside quotes.
+function readQuoted(sql, i, quote) {
+    let j = i + 1;
+    while (j < sql.length) {
+        const char = sql[j];
+        if (char === '\\') {
+            j += 2;
+            continue;
+        }
+        if (char === quote) {
+            if (sql[j + 1] === quote) {
+                j += 2;
+                continue;
+            }
+            return { end: j + 1, terminated: true };
+        }
+        j += 1;
+    }
+    return { end: sql.length, terminated: false };
+}
+// `$tag$ … $tag$` (including `$$ … $$`), where the tag is `[A-Za-z0-9_]*`. It
+// is only a literal when the same tag closes it later; otherwise ClickHouse
+// reads the `$` as part of an identifier or as punctuation.
+function readHeredoc(sql, i) {
+    const tagEnd = skipWhile(sql, i + 1, HEREDOC_TAG);
+    if (sql[tagEnd] !== '$')
+        return undefined;
+    const tag = sql.slice(i, tagEnd + 1);
+    const close = sql.indexOf(tag, tagEnd + 1);
+    return close === -1 ? undefined : close + tag.length;
+}
+// Resolves a doubled delimiter and backslash escapes to the escaped character,
+// which is exact for the quotes and backslashes names and literals contain.
+function unescapeQuoted(text) {
+    const quote = text[0] ?? '';
+    const body = text.slice(1, -1);
+    let out = '';
+    for (let i = 0; i < body.length; i += 1) {
+        const char = body[i] ?? '';
+        if (char === '\\' && i + 1 < body.length) {
+            out += body[i + 1];
+            i += 1;
+            continue;
+        }
+        if (char === quote && body[i + 1] === quote) {
+            out += quote;
+            i += 1;
+            continue;
+        }
+        out += char;
+    }
+    return out;
+}
+function complete(kind, end) {
+    return { kind, end, terminated: true };
+}
+function skipWhile(sql, i, pattern) {
+    let j = i;
+    while (j < sql.length && pattern.test(sql[j] ?? ''))
+        j += 1;
+    return j;
+}
+function lineEnd(sql, i) {
+    const newline = sql.indexOf('\n', i);
+    return newline === -1 ? sql.length : newline;
+}
+//# sourceMappingURL=sql-lexer.js.map

package/dist/sql-lexer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"sql-lexer.js","sourceRoot":"","sources":["../src/sql-lexer.ts"],"names":[],"mappings":"AAqCA,MAAM,UAAU,GAAG,eAAe,CAAA;AAClC,MAAM,gBAAgB,GAAG,WAAW,CAAA;AACpC,MAAM,eAAe,GAAG,eAAe,CAAA;AACvC,MAAM,SAAS,GAAG,cAAc,CAAA;AAChC,MAAM,IAAI,GAAG,iBAAiB,CAAA;AAC9B,MAAM,KAAK,GAAG,OAAO,CAAA;AACrB,MAAM,SAAS,GAAG,aAAa,CAAA;AAC/B,MAAM,WAAW,GAAG,cAAc,CAAA;AAClC,MAAM,QAAQ,GAAG,iBAAiB,CAAA;AAClC,MAAM,YAAY,GAAG,iBAAiB,CAAA;AAEtC,MAAM,UAAU,WAAW,CAAC,GAAW;IACrC,MAAM,MAAM,GAAe,EAAE,CAAA;IAC7B,IAAI,QAA8B,CAAA;IAClC,IAAI,CAAC,GAAG,CAAC,CAAA;IACT,OAAO,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;QACtB,MAAM,EAAE,IAAI,EAAE,GAAG,EAAE,UAAU,EAAE,GAAG,SAAS,CAAC,GAAG,EAAE,CAAC,EAAE,QAAQ,CAAC,CAAA;QAC7D,MAAM,KAAK,GAAa,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,GAAG,EAAE,UAAU,EAAE,CAAA;QACpF,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;QAClB,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;YAAE,QAAQ,GAAG,KAAK,CAAA;QACtC,CAAC,GAAG,GAAG,CAAA;IACT,CAAC;IACD,OAAO,MAAM,CAAA;AACf,CAAC;AAED,gFAAgF;AAChF,MAAM,UAAU,QAAQ,CAAC,KAAe;IACtC,OAAO,KAAK,CAAC,IAAI,KAAK,YAAY,IAAI,KAAK,CAAC,IAAI,KAAK,cAAc,IAAI,KAAK,CAAC,IAAI,KAAK,eAAe,CAAA;AACvG,CAAC;AAED,8EAA8E;AAC9E,MAAM,UAAU,cAAc,CAAC,KAAe;IAC5C,IAAI,KAAK,CAAC,IAAI,KAAK,YAAY;QAAE,OAAO,KAAK,CAAC,IAAI,CAAA;IAClD,IAAI,KAAK,CAAC,IAAI,KAAK,mBAAmB,IAAI,CAAC,KAAK,CAAC,UAAU;QAAE,OAAO,SAAS,CAAA;IAC7E,OAAO,cAAc,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;AACnC,CAAC;AAED,mFAAmF;AACnF,MAAM,UAAU,kBAAkB,CAAC,KAAe;IAChD,IAAI,KAAK,CAAC,IAAI,KAAK,QAAQ,IAAI,CAAC,KAAK,CAAC,UAAU;QAAE,OAAO,SAAS,CAAA;IAClE,IAAI,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;QAC/B,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,CAAA;QAChD,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,EAAE,KAAK,CAAC,IAAI,CAAC,MAAM,GAAG,SAAS,CAAC,CAAA;IACnE,CAAC;IACD,OAAO,cAAc,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;AACnC,CAAC;AAED,8EAA8E;AAC9E,8CAA8C;AAC9C,SAAS,SAAS,CAAC,GAAW,EAAE,CAAS,EAAE,QAA8B;IACvE,MAAM,IAAI,GAAG,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;IACzB,MAAM,IAAI,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;IAC7B,IAAI,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,QAAQ,CAAC,YAAY,EAAE,SAAS,CAAC,GAAG,EAAE,CAAC,EAAE,UAAU,CAAC,CAAC,CAAA;IACvF,wFAAwF;IACxF,IACE,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC;QAC9B,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC;QAC9B,CAAC,IAAI,KAAK,GAAG,IAAI,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,CAAC,CAAC,EAChD,CAAC;QACD,OAAO,QAAQ,CAAC,cAAc,EAAE,OAAO,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAA;IAClD,CAAC;IACD,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG;QAAE,OAAO,gBAAgB,CAAC,GAAG,EAAE,CAAC,CAAC,CAAA;IACjE,IAAI,IAAI,KAAK,GAAG;QAAE,OAAO,EAAE,IAAI,EAAE,QAAQ,EAAE,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC,EAAE,GAAG,CAAC,EAAE,CAAA;IACvE,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG;QAAE,OAAO,EAAE,IAAI,EAAE,mBAAmB,EAAE,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC,EAAE,IAAI,CAAC,EAAE,CAAA;IACnG,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;QACjB,MAAM,UAAU,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,CAAA;QACtC,IAAI,UAAU,KAAK,SAAS;YAAE,OAAO,QAAQ,CAAC,QAAQ,EAAE,UAAU,CAAC,CAAA;QACnE,IAAI,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC;YAAE,OAAO,QAAQ,CAAC,YAAY,EAAE,SAAS,CAAC,GAAG,EAAE,CAAC,EAAE,eAAe,CAAC,CAAC,CAAA;IAC7F,CAAC;IACD,IAAI,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,gBAAgB,CAAC,GAAG,EAAE,CAAC,EAAE,QAAQ,EAAE,IAAI,KAAK,GAAG,CAAC,CAAA;IAC7E,IAAI,IAAI,KAAK,GAAG,IAAI,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,QAAQ,CAAC;QAAE,OAAO,QAAQ,CAAC,QAAQ,EAAE,WAAW,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAA;IAC9G,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,QAAQ,CAAC,YAAY,EAAE,SAAS,CAAC,GAAG,EAAE,CAAC,EAAE,eAAe,CAAC,CAAC,CAAA;IAClG,MAAM,SAAS,GAAG,GAAG,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,CAAA;IACzC,OAAO,QAAQ,CAAC,aAAa,EAAE,CAAC,GAAG,MAAM,CAAC,aAAa,CAAC,SAAS,CAAC,CAAC,MAAM,CAAC,CAAA;AAC5E,CAAC;AAED,uEAAuE;AACvE,8EAA8E;AAC9E,yEAAyE;AACzE,gFAAgF;AAChF,2EAA2E;AAC3E,SAAS,gBAAgB,CAAC,GAAW,EAAE,CAAS,EAAE,QAAiB;IACjE,MAAM,UAAU,GAAG,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,iBAAiB,CAAC,GAAG,EAAE,CAAC,CAAC,CAAA;IAClF,MAAM,GAAG,GAAG,SAAS,CAAC,GAAG,EAAE,UAAU,EAAE,SAAS,CAAC,CAAA;IACjD,MAAM,MAAM,GAAG,GAAG,GAAG,UAAU,IAAI,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAA;IAC/D,OAAO,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAA;AACxD,CAAC;AAED,yEAAyE;AACzE,iFAAiF;AACjF,SAAS,iBAAiB,CAAC,GAAW,EAAE,CAAS;IAC/C,MAAM,MAAM,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAA;IAClC,MAAM,GAAG,GAAG,mBAAmB,CAAC,IAAI,CAAC,MAAM,CAAC,CAAA;IAC5C,MAAM,KAAK,GAAG,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAA;IACrC,IAAI,CAAC,GAAG,SAAS,CAAC,GAAG,EAAE,GAAG,IAAI,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,CAAC,CAAA;IAC3E,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,GAAG,IAAI,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QAAE,CAAC,GAAG,SAAS,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,EAAE,KAAK,CAAC,CAAA;IACpF,MAAM,QAAQ,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,YAAY,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAA;IAC1E,OAAO,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,EAAE,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;AACzE,CAAC;AAED,8DAA8D;AAC9D,SAAS,SAAS,CAAC,GAAW,EAAE,CAAS,EAAE,KAAa;IACtD,IAAI,CAAC,GAAG,CAAC,CAAA;IACT,OAAO,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;QACtB,MAAM,IAAI,GAAG,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;QACzB,MAAM,SAAS,GAAG,IAAI,KAAK,GAAG,IAAI,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;QACvE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,SAAS;YAAE,MAAK;QAC1C,CAAC,IAAI,CAAC,CAAA;IACR,CAAC;IACD,OAAO,CAAC,CAAA;AACV,CAAC;AAED,uCAAuC;AACvC,SAAS,WAAW,CAAC,GAAW,EAAE,CAAS;IACzC,MAAM,CAAC,GAAG,SAAS,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,EAAE,KAAK,CAAC,CAAA;IACtC,MAAM,QAAQ,GAAG,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAA;IACnD,OAAO,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,EAAE,CAAC,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAA;AACzE,CAAC;AAED,4EAA4E;AAC5E,SAAS,WAAW,CAAC,KAA2B;IAC9C,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAA;IACrC,OAAO,CACL,KAAK,CAAC,IAAI,KAAK,YAAY;QAC3B,KAAK,CAAC,IAAI,KAAK,mBAAmB;QAClC,KAAK,CAAC,IAAI,KAAK,QAAQ;QACvB,KAAK,CAAC,IAAI,KAAK,GAAG;QAClB,KAAK,CAAC,IAAI,KAAK,GAAG,CACnB,CAAA;AACH,CAAC;AAED,sEAAsE;AACtE,SAAS,gBAAgB,CAAC,GAAW,EAAE,CAAS;IAC9C,IAAI,KAAK,GAAG,CAAC,CAAA;IACb,IAAI,CAAC,GAAG,CAAC,CAAA;IACT,OAAO,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;QACtB,IAAI,GAAG,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,CAAC,EAAE,CAAC;YAC5B,KAAK,IAAI,CAAC,CAAA;YACV,CAAC,IAAI,CAAC,CAAA;QACR,CAAC;aAAM,IAAI,GAAG,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,CAAC,EAAE,CAAC;YACnC,KAAK,IAAI,CAAC,CAAA;YACV,CAAC,IAAI,CAAC,CAAA;YACN,IAAI,KAAK,KAAK,CAAC;gBAAE,OAAO,QAAQ,CAAC,eAAe,EAAE,CAAC,CAAC,CAAA;QACtD,CAAC;aAAM,CAAC;YACN,CAAC,IAAI,CAAC,CAAA;QACR,CAAC;IACH,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,eAAe,EAAE,GAAG,EAAE,GAAG,CAAC,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,CAAA;AACtE,CAAC;AAED,0EAA0E;AAC1E,SAAS,UAAU,CAAC,GAAW,EAAE,CAAS,EAAE,KAAa;IACvD,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAA;IACb,OAAO,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;QACtB,MAAM,IAAI,GAAG,GAAG,CAAC,CAAC,CAAC,CAAA;QACnB,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;YAClB,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,KAAK,KAAK,EAAE,CAAC;YACnB,IAAI,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,KAAK,EAAE,CAAC;gBACzB,CAAC,IAAI,CAAC,CAAA;gBACN,SAAQ;YACV,CAAC;YACD,OAAO,EAAE,GAAG,EAAE,CAAC,GAAG,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,CAAA;QACzC,CAAC;QACD,CAAC,IAAI,CAAC,CAAA;IACR,CAAC;IACD,OAAO,EAAE,GAAG,EAAE,GAAG,CAAC,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,CAAA;AAC/C,CAAC;AAED,8EAA8E;AAC9E,4EAA4E;AAC5E,4DAA4D;AAC5D,SAAS,WAAW,CAAC,GAAW,EAAE,CAAS;IACzC,MAAM,MAAM,GAAG,SAAS,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,EAAE,WAAW,CAAC,CAAA;IACjD,IAAI,GAAG,CAAC,MAAM,CAAC,KAAK,GAAG;QAAE,OAAO,SAAS,CAAA;IACzC,MAAM,GAAG,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,MAAM,GAAG,CAAC,CAAC,CAAA;IACpC,MAAM,KAAK,GAAG,GAAG,CAAC,OAAO,CAAC,GAAG,EAAE,MAAM,GAAG,CAAC,CAAC,CAAA;IAC1C,OAAO,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,GAAG,GAAG,CAAC,MAAM,CAAA;AACtD,CAAC;AAED,+EAA+E;AAC/E,4EAA4E;AAC5E,SAAS,cAAc,CAAC,IAAY;IAClC,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;IAC3B,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IAC9B,IAAI,GAAG,GAAG,EAAE,CAAA;IACZ,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACxC,MAAM,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAA;QAC1B,IAAI,IAAI,KAAK,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;YACzC,GAAG,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;YAClB,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,KAAK,EAAE,CAAC;YAC5C,GAAG,IAAI,KAAK,CAAA;YACZ,CAAC,IAAI,CAAC,CAAA;YACN,SAAQ;QACV,CAAC;QACD,GAAG,IAAI,IAAI,CAAA;IACb,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,SAAS,QAAQ,CAAC,IAAkB,EAAE,GAAW;IAC/C,OAAO,EAAE,IAAI,EAAE,GAAG,EAAE,UAAU,EAAE,IAAI,EAAE,CAAA;AACxC,CAAC;AAED,SAAS,SAAS,CAAC,GAAW,EAAE,CAAS,EAAE,OAAe;IACxD,IAAI,CAAC,GAAG,CAAC,CAAA;IACT,OAAO,CAAC,GAAG,GAAG,CAAC,MAAM,IAAI,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;QAAE,CAAC,IAAI,CAAC,CAAA;IAC3D,OAAO,CAAC,CAAA;AACV,CAAC;AAED,SAAS,OAAO,CAAC,GAAW,EAAE,CAAS;IACrC,MAAM,OAAO,GAAG,GAAG,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC,CAAA;IACpC,OAAO,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,CAAA;AAC9C,CAAC"}

package/dist/sql-normalizer.d.ts

@@ -1,3 +1,51 @@
+/**
+ * Compare expression tokens while preserving quoted values and identifier case.
+ * SQL the lexer cannot read (e.g. an unterminated quote) falls back to its
+ * normalizeSQLFragment form, encoded as a JSON string so it never equals a token
+ * list.
+ */
+export declare function sqlExpressionFingerprint(value: string): string;
+/**
+ * Whether `expression` is the default ClickHouse stores for a bare EPHEMERAL
+ * column. ClickHouse synthesizes defaultValueOfTypeName('<type as written>');
+ * the literal keeps alias spellings (BIGINT, TEXT) that system.columns.type
+ * canonicalizes, so any single string literal counts.
+ */
+export declare function isSyntheticEphemeralDefault(expression: string): boolean;
+/**
+ * Canonical one-line form of a SQL fragment: a view or materialized view `as`,
+ * `partitionBy`, `ttl`, a skip index expression, a projection, a dictionary
+ * `source`/`layout`/`lifetime`, or the same clauses read back from ClickHouse.
+ *
+ * Comments are dropped and every run of whitespace becomes one space, inside
+ * string literals too. Comments must go before the newlines do: a `--`, `//`,
+ * `# ` or `#!` comment ends at its newline, so on a single line it would swallow
+ * the rest of the fragment and the `;` the renderer appends after it (#232).
+ * Comment syntax is ClickHouse's, from the shared lexer: markers inside string
+ * literals, quoted identifiers and heredocs are text, and an unterminated block
+ * comment or string is kept as written for ClickHouse to report.
+ *
+ * A dropped comment leaves one space, so the tokens around it never fuse: a
+ * minus, an empty block comment and a minus become `- -`, not a `--` comment.
+ * Text without comments or a stray `#` normalizes exactly as a plain
+ * whitespace collapse would, and the result is stable: normalizing it again
+ * changes nothing.
+ */
 export declare function normalizeSQLFragment(value: string): string;
+/**
+ * SQL text without its comments, trimmed. Unlike normalizeSQLFragment, all
+ * other text stays as written, including whitespace inside string literals.
+ * chkit renders a column default expression on one line of a CREATE or ALTER
+ * statement, where a `--` comment would swallow the `,`, `COMMENT`, `CODEC` or
+ * `;` after it (#234).
+ *
+ * A run of whitespace and comments that holds a comment becomes one space, so
+ * the tokens around it never fuse, or a newline after a stray `#`, for the
+ * reason normalizeSQLFragment gives. A stray `#` at the end is followed by a
+ * newline too: the ` COMMENT` or ` CODEC` rendered after it would otherwise
+ * turn it into a `# ` comment that hides them. An unterminated block comment
+ * or string is kept as written for ClickHouse to report.
+ */
+export declare function stripSQLComments(value: string): string;
 export declare function normalizeEngine(engine: string): string;
 //# sourceMappingURL=sql-normalizer.d.ts.map

package/dist/sql-normalizer.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"sql-normalizer.d.ts","sourceRoot":"","sources":["../src/sql-normalizer.ts"],"names":[],"mappings":"AAAA,wBAAgB,oBAAoB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAE1D;AAED,wBAAgB,eAAe,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,CAMtD"}
+{"version":3,"file":"sql-normalizer.d.ts","sourceRoot":"","sources":["../src/sql-normalizer.ts"],"names":[],"mappings":"AAMA;;;;;GAKG;AACH,wBAAgB,wBAAwB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAM9D;AAED;;;;;GAKG;AACH,wBAAgB,2BAA2B,CAAC,UAAU,EAAE,MAAM,GAAG,OAAO,CAEvE;AAED;;;;;;;;;;;;;;;;;;GAkBG;AACH,wBAAgB,oBAAoB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAmB1D;AAED;;;;;;;;;;;;;GAaG;AACH,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAsBtD;AAED,wBAAgB,eAAe,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,CAOtD"}

package/dist/sql-normalizer.js

@@ -1,11 +1,128 @@
+import { isKafkaEngine, normalizeKafkaEngine } from './kafka.js';
+import { tokenizeSQL } from './sql-lexer.js';
+import { textExpressionFingerprint, textSQLFingerprint } from './text-index-sql.js';
+const SYNTHETIC_EPHEMERAL_DEFAULT = /^defaultValueOfTypeName\s*\(\s*'(?:[^'\\]|\\.|'')*'\s*\)$/s;
+/**
+ * Compare expression tokens while preserving quoted values and identifier case.
+ * SQL the lexer cannot read (e.g. an unterminated quote) falls back to its
+ * normalizeSQLFragment form, encoded as a JSON string so it never equals a token
+ * list.
+ */
+export function sqlExpressionFingerprint(value) {
+    try {
+        return textSQLFingerprint(textExpressionFingerprint(value));
+    }
+    catch {
+        return JSON.stringify(normalizeSQLFragment(value));
+    }
+}
+/**
+ * Whether `expression` is the default ClickHouse stores for a bare EPHEMERAL
+ * column. ClickHouse synthesizes defaultValueOfTypeName('<type as written>');
+ * the literal keeps alias spellings (BIGINT, TEXT) that system.columns.type
+ * canonicalizes, so any single string literal counts.
+ */
+export function isSyntheticEphemeralDefault(expression) {
+    return SYNTHETIC_EPHEMERAL_DEFAULT.test(expression.trim());
+}
+/**
+ * Canonical one-line form of a SQL fragment: a view or materialized view `as`,
+ * `partitionBy`, `ttl`, a skip index expression, a projection, a dictionary
+ * `source`/`layout`/`lifetime`, or the same clauses read back from ClickHouse.
+ *
+ * Comments are dropped and every run of whitespace becomes one space, inside
+ * string literals too. Comments must go before the newlines do: a `--`, `//`,
+ * `# ` or `#!` comment ends at its newline, so on a single line it would swallow
+ * the rest of the fragment and the `;` the renderer appends after it (#232).
+ * Comment syntax is ClickHouse's, from the shared lexer: markers inside string
+ * literals, quoted identifiers and heredocs are text, and an unterminated block
+ * comment or string is kept as written for ClickHouse to report.
+ *
+ * A dropped comment leaves one space, so the tokens around it never fuse: a
+ * minus, an empty block comment and a minus become `- -`, not a `--` comment.
+ * Text without comments or a stray `#` normalizes exactly as a plain
+ * whitespace collapse would, and the result is stable: normalizing it again
+ * changes nothing.
+ */
 export function normalizeSQLFragment(value) {
-    return value.replace(/\s+/g, ' ').trim();
+    let out = '';
+    let separator = '';
+    let afterBareHash = false;
+    for (const token of tokenizeSQL(value)) {
+        if (separatesTokens(token)) {
+            // `#` not followed by a space or `!` is a syntax error in ClickHouse.
+            // A space after it would turn it into a `# ` comment that hides the rest
+            // of the fragment, so keep the error visible with a newline instead.
+            separator = afterBareHash ? '\n' : ' ';
+            continue;
+        }
+        if (out !== '')
+            out += separator;
+        out += token.text.replace(/\s+/g, ' ');
+        separator = '';
+        afterBareHash = token.kind === 'punctuation' && token.text === '#';
+    }
+    // An unterminated string or block comment at the end can still end in a space.
+    return out.trim();
+}
+/**
+ * SQL text without its comments, trimmed. Unlike normalizeSQLFragment, all
+ * other text stays as written, including whitespace inside string literals.
+ * chkit renders a column default expression on one line of a CREATE or ALTER
+ * statement, where a `--` comment would swallow the `,`, `COMMENT`, `CODEC` or
+ * `;` after it (#234).
+ *
+ * A run of whitespace and comments that holds a comment becomes one space, so
+ * the tokens around it never fuse, or a newline after a stray `#`, for the
+ * reason normalizeSQLFragment gives. A stray `#` at the end is followed by a
+ * newline too: the ` COMMENT` or ` CODEC` rendered after it would otherwise
+ * turn it into a `# ` comment that hides them. An unterminated block comment
+ * or string is kept as written for ClickHouse to report.
+ */
+export function stripSQLComments(value) {
+    let out = '';
+    let gap = '';
+    let gapHasComment = false;
+    let afterBareHash = false;
+    let endsWithBareHash = false;
+    for (const token of tokenizeSQL(value)) {
+        const comment = isClosedComment(token);
+        if (comment || token.kind === 'whitespace') {
+            gap += token.text;
+            gapHasComment = gapHasComment || comment;
+            continue;
+        }
+        out += gapHasComment ? (afterBareHash ? '\n' : ' ') : gap;
+        out += token.text;
+        gap = '';
+        gapHasComment = false;
+        afterBareHash = token.kind === 'punctuation' && token.text === '#';
+        // trim() also drops Unicode whitespace, which the lexer reads as punctuation.
+        if (!/^\s$/.test(token.text))
+            endsWithBareHash = afterBareHash;
+    }
+    return endsWithBareHash ? `${out.trim()}\n` : out.trim();
 }
 export function normalizeEngine(engine) {
+    if (isKafkaEngine(engine))
+        return normalizeKafkaEngine(engine);
     let normalized = engine.trim().replace(/^Shared/, '');
     if (!normalized.includes('(')) {
         normalized += '()';
     }
     return normalized;
 }
+// Whitespace and closed comments. The lexer reads Unicode whitespace such as a
+// no-break space as punctuation; ClickHouse skips it like a space, and chkit
+// has always collapsed it like one.
+function separatesTokens(token) {
+    if (token.kind === 'whitespace')
+        return true;
+    if (token.kind === 'line_comment' || token.kind === 'block_comment')
+        return token.terminated;
+    return token.kind === 'punctuation' && /^\s$/.test(token.text);
+}
+function isClosedComment(token) {
+    return (token.kind === 'line_comment' || token.kind === 'block_comment') && token.terminated;
+}
 //# sourceMappingURL=sql-normalizer.js.map

package/dist/sql-normalizer.js.map

@@ -1 +1 @@
-{"version":3,"file":"sql-normalizer.js","sourceRoot":"","sources":["../src/sql-normalizer.ts"],"names":[],"mappings":"AAAA,MAAM,UAAU,oBAAoB,CAAC,KAAa;IAChD,OAAO,KAAK,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAA;AAC1C,CAAC;AAED,MAAM,UAAU,eAAe,CAAC,MAAc;IAC5C,IAAI,UAAU,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAA;IACrD,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC9B,UAAU,IAAI,IAAI,CAAA;IACpB,CAAC;IACD,OAAO,UAAU,CAAA;AACnB,CAAC"}
+{"version":3,"file":"sql-normalizer.js","sourceRoot":"","sources":["../src/sql-normalizer.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,oBAAoB,EAAE,MAAM,YAAY,CAAA;AAChE,OAAO,EAAiB,WAAW,EAAE,MAAM,gBAAgB,CAAA;AAC3D,OAAO,EAAE,yBAAyB,EAAE,kBAAkB,EAAE,MAAM,qBAAqB,CAAA;AAEnF,MAAM,2BAA2B,GAAG,4DAA4D,CAAA;AAEhG;;;;;GAKG;AACH,MAAM,UAAU,wBAAwB,CAAC,KAAa;IACpD,IAAI,CAAC;QACH,OAAO,kBAAkB,CAAC,yBAAyB,CAAC,KAAK,CAAC,CAAC,CAAA;IAC7D,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC,SAAS,CAAC,oBAAoB,CAAC,KAAK,CAAC,CAAC,CAAA;IACpD,CAAC;AACH,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,2BAA2B,CAAC,UAAkB;IAC5D,OAAO,2BAA2B,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,EAAE,CAAC,CAAA;AAC5D,CAAC;AAED;;;;;;;;;;;;;;;;;;GAkBG;AACH,MAAM,UAAU,oBAAoB,CAAC,KAAa;IAChD,IAAI,GAAG,GAAG,EAAE,CAAA;IACZ,IAAI,SAAS,GAAG,EAAE,CAAA;IAClB,IAAI,aAAa,GAAG,KAAK,CAAA;IACzB,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,KAAK,CAAC,EAAE,CAAC;QACvC,IAAI,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC;YAC3B,sEAAsE;YACtE,yEAAyE;YACzE,qEAAqE;YACrE,SAAS,GAAG,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAA;YACtC,SAAQ;QACV,CAAC;QACD,IAAI,GAAG,KAAK,EAAE;YAAE,GAAG,IAAI,SAAS,CAAA;QAChC,GAAG,IAAI,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAA;QACtC,SAAS,GAAG,EAAE,CAAA;QACd,aAAa,GAAG,KAAK,CAAC,IAAI,KAAK,aAAa,IAAI,KAAK,CAAC,IAAI,KAAK,GAAG,CAAA;IACpE,CAAC;IACD,+EAA+E;IAC/E,OAAO,GAAG,CAAC,IAAI,EAAE,CAAA;AACnB,CAAC;AAED;;;;;;;;;;;;;GAaG;AACH,MAAM,UAAU,gBAAgB,CAAC,KAAa;IAC5C,IAAI,GAAG,GAAG,EAAE,CAAA;IACZ,IAAI,GAAG,GAAG,EAAE,CAAA;IACZ,IAAI,aAAa,GAAG,KAAK,CAAA;IACzB,IAAI,aAAa,GAAG,KAAK,CAAA;IACzB,IAAI,gBAAgB,GAAG,KAAK,CAAA;IAC5B,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,KAAK,CAAC,EAAE,CAAC;QACvC,MAAM,OAAO,GAAG,eAAe,CAAC,KAAK,CAAC,CAAA;QACtC,IAAI,OAAO,IAAI,KAAK,CAAC,IAAI,KAAK,YAAY,EAAE,CAAC;YAC3C,GAAG,IAAI,KAAK,CAAC,IAAI,CAAA;YACjB,aAAa,GAAG,aAAa,IAAI,OAAO,CAAA;YACxC,SAAQ;QACV,CAAC;QACD,GAAG,IAAI,aAAa,CAAC,CAAC,CAAC,CAAC,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAA;QACzD,GAAG,IAAI,KAAK,CAAC,IAAI,CAAA;QACjB,GAAG,GAAG,EAAE,CAAA;QACR,aAAa,GAAG,KAAK,CAAA;QACrB,aAAa,GAAG,KAAK,CAAC,IAAI,KAAK,aAAa,IAAI,KAAK,CAAC,IAAI,KAAK,GAAG,CAAA;QAClE,8EAA8E;QAC9E,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC;YAAE,gBAAgB,GAAG,aAAa,CAAA;IAChE,CAAC;IACD,OAAO,gBAAgB,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,IAAI,EAAE,CAAA;AAC1D,CAAC;AAED,MAAM,UAAU,eAAe,CAAC,MAAc;IAC5C,IAAI,aAAa,CAAC,MAAM,CAAC;QAAE,OAAO,oBAAoB,CAAC,MAAM,CAAC,CAAA;IAC9D,IAAI,UAAU,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAA;IACrD,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC9B,UAAU,IAAI,IAAI,CAAA;IACpB,CAAC;IACD,OAAO,UAAU,CAAA;AACnB,CAAC;AAED,+EAA+E;AAC/E,6EAA6E;AAC7E,oCAAoC;AACpC,SAAS,eAAe,CAAC,KAAe;IACtC,IAAI,KAAK,CAAC,IAAI,KAAK,YAAY;QAAE,OAAO,IAAI,CAAA;IAC5C,IAAI,KAAK,CAAC,IAAI,KAAK,cAAc,IAAI,KAAK,CAAC,IAAI,KAAK,eAAe;QAAE,OAAO,KAAK,CAAC,UAAU,CAAA;IAC5F,OAAO,KAAK,CAAC,IAAI,KAAK,aAAa,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;AAChE,CAAC;AAED,SAAS,eAAe,CAAC,KAAe;IACtC,OAAO,CAAC,KAAK,CAAC,IAAI,KAAK,cAAc,IAAI,KAAK,CAAC,IAAI,KAAK,eAAe,CAAC,IAAI,KAAK,CAAC,UAAU,CAAA;AAC9F,CAAC"}

package/dist/sql-scan.d.ts

@@ -0,0 +1,20 @@
+type QuoteChar = "'" | '"' | '`';
+export declare function isQuoteChar(char: string | undefined): char is QuoteChar;
+export declare function findQuoteEnd(sql: string, start: number, quote: string): number;
+/**
+ * Peel one layer of wrapping parentheses, but only when the leading `(` closes
+ * at the very end — so `(a, b)` becomes `a, b` while `(a), (b)` is left intact.
+ */
+export declare function stripWrappingParens(input: string): string;
+/**
+ * Index of the `)` that matches the `(` at `openIndex`, or undefined when the
+ * parens are unbalanced.
+ */
+export declare function findMatchingParen(input: string, openIndex: number): number | undefined;
+/** Find a clause/delimiter outside quoted strings, identifiers and parentheses. */
+export declare function findTopLevelSQLPattern(sql: string, pattern: RegExp): {
+    index: number;
+    length: number;
+} | undefined;
+export {};
+//# sourceMappingURL=sql-scan.d.ts.map

package/dist/sql-scan.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"sql-scan.d.ts","sourceRoot":"","sources":["../src/sql-scan.ts"],"names":[],"mappings":"AAQA,KAAK,SAAS,GAAG,GAAG,GAAG,GAAG,GAAG,GAAG,CAAA;AAEhC,wBAAgB,WAAW,CAAC,IAAI,EAAE,MAAM,GAAG,SAAS,GAAG,IAAI,IAAI,SAAS,CAEvE;AAID,wBAAgB,YAAY,CAAC,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,GAAG,MAAM,CAS9E;AAED;;;GAGG;AACH,wBAAgB,mBAAmB,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,CAIzD;AAED;;;GAGG;AACH,wBAAgB,iBAAiB,CAAC,KAAK,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAYtF;AAED,mFAAmF;AACnF,wBAAgB,sBAAsB,CACpC,GAAG,EAAE,MAAM,EACX,OAAO,EAAE,MAAM,GACd;IAAE,KAAK,EAAE,MAAM,CAAC;IAAC,MAAM,EAAE,MAAM,CAAA;CAAE,GAAG,SAAS,CAyB/C"}

package/dist/sql-scan.js

@@ -0,0 +1,82 @@
+// Quote-aware SQL scanning primitives. Characters inside a single-quoted
+// string, a double-quoted identifier, or a backtick-quoted identifier are
+// literal text, not structure — parens and commas there must be ignored.
+//
+// Every scanner skips quoted regions through `findQuoteEnd`, so the escaping
+// rule lives in exactly one place and the copies cannot drift apart again
+// (#197).
+export function isQuoteChar(char) {
+    return char === "'" || char === '"' || char === '`';
+}
+// Index of the quote closing the one at `start`; honours backslash escapes and
+// doubled quotes. Unterminated input runs to the end of the string.
+export function findQuoteEnd(sql, start, quote) {
+    for (let i = start + 1; i < sql.length; i += 1) {
+        if (sql[i] === '\\')
+            i += 1;
+        else if (sql[i] === quote) {
+            if (sql[i + 1] === quote)
+                i += 1;
+            else
+                return i;
+        }
+    }
+    return sql.length;
+}
+/**
+ * Peel one layer of wrapping parentheses, but only when the leading `(` closes
+ * at the very end — so `(a, b)` becomes `a, b` while `(a), (b)` is left intact.
+ */
+export function stripWrappingParens(input) {
+    if (!input.startsWith('(') || !input.endsWith(')'))
+        return input;
+    const close = findMatchingParen(input, 0);
+    return close === input.length - 1 ? input.slice(1, -1).trim() : input;
+}
+/**
+ * Index of the `)` that matches the `(` at `openIndex`, or undefined when the
+ * parens are unbalanced.
+ */
+export function findMatchingParen(input, openIndex) {
+    let depth = 0;
+    for (let i = openIndex; i < input.length; i += 1) {
+        const char = input[i];
+        if (isQuoteChar(char))
+            i = findQuoteEnd(input, i, char);
+        else if (char === '(')
+            depth += 1;
+        else if (char === ')') {
+            depth -= 1;
+            if (depth === 0)
+                return i;
+        }
+    }
+    return undefined;
+}
+/** Find a clause/delimiter outside quoted strings, identifiers and parentheses. */
+export function findTopLevelSQLPattern(sql, pattern) {
+    let depth = 0;
+    for (let i = 0; i < sql.length; i += 1) {
+        const char = sql[i];
+        if (isQuoteChar(char)) {
+            i = findQuoteEnd(sql, i, char);
+            continue;
+        }
+        if (char === '(') {
+            depth += 1;
+            continue;
+        }
+        if (char === ')') {
+            depth -= 1;
+            continue;
+        }
+        if (depth !== 0 ||
+            (/[A-Za-z_]/.test(char ?? '') && i > 0 && /[A-Za-z0-9_]/.test(sql[i - 1] ?? '')))
+            continue;
+        const match = sql.slice(i).match(pattern);
+        if (match?.index === 0)
+            return { index: i, length: match[0].length };
+    }
+    return undefined;
+}
+//# sourceMappingURL=sql-scan.js.map

package/dist/sql-scan.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"sql-scan.js","sourceRoot":"","sources":["../src/sql-scan.ts"],"names":[],"mappings":"AAAA,yEAAyE;AACzE,0EAA0E;AAC1E,yEAAyE;AACzE,EAAE;AACF,6EAA6E;AAC7E,0EAA0E;AAC1E,UAAU;AAIV,MAAM,UAAU,WAAW,CAAC,IAAwB;IAClD,OAAO,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,GAAG,CAAA;AACrD,CAAC;AAED,+EAA+E;AAC/E,oEAAoE;AACpE,MAAM,UAAU,YAAY,CAAC,GAAW,EAAE,KAAa,EAAE,KAAa;IACpE,KAAK,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QAC/C,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI;YAAE,CAAC,IAAI,CAAC,CAAA;aACtB,IAAI,GAAG,CAAC,CAAC,CAAC,KAAK,KAAK,EAAE,CAAC;YAC1B,IAAI,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,KAAK;gBAAE,CAAC,IAAI,CAAC,CAAA;;gBAC3B,OAAO,CAAC,CAAA;QACf,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAC,MAAM,CAAA;AACnB,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,mBAAmB,CAAC,KAAa;IAC/C,IAAI,CAAC,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC;QAAE,OAAO,KAAK,CAAA;IAChE,MAAM,KAAK,GAAG,iBAAiB,CAAC,KAAK,EAAE,CAAC,CAAC,CAAA;IACzC,OAAO,KAAK,KAAK,KAAK,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,KAAK,CAAA;AACvE,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,iBAAiB,CAAC,KAAa,EAAE,SAAiB;IAChE,IAAI,KAAK,GAAG,CAAC,CAAA;IACb,KAAK,IAAI,CAAC,GAAG,SAAS,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACjD,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,CAAA;QACrB,IAAI,WAAW,CAAC,IAAI,CAAC;YAAE,CAAC,GAAG,YAAY,CAAC,KAAK,EAAE,CAAC,EAAE,IAAI,CAAC,CAAA;aAClD,IAAI,IAAI,KAAK,GAAG;YAAE,KAAK,IAAI,CAAC,CAAA;aAC5B,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACtB,KAAK,IAAI,CAAC,CAAA;YACV,IAAI,KAAK,KAAK,CAAC;gBAAE,OAAO,CAAC,CAAA;QAC3B,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC;AAED,mFAAmF;AACnF,MAAM,UAAU,sBAAsB,CACpC,GAAW,EACX,OAAe;IAEf,IAAI,KAAK,GAAG,CAAC,CAAA;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,GAAG,CAAC,CAAC,CAAC,CAAA;QACnB,IAAI,WAAW,CAAC,IAAI,CAAC,EAAE,CAAC;YACtB,CAAC,GAAG,YAAY,CAAC,GAAG,EAAE,CAAC,EAAE,IAAI,CAAC,CAAA;YAC9B,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,KAAK,IAAI,CAAC,CAAA;YACV,SAAQ;QACV,CAAC;QACD,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;YACjB,KAAK,IAAI,CAAC,CAAA;YACV,SAAQ;QACV,CAAC;QACD,IACE,KAAK,KAAK,CAAC;YACX,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,IAAI,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;YAEhF,SAAQ;QACV,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAA;QACzC,IAAI,KAAK,EAAE,KAAK,KAAK,CAAC;YAAE,OAAO,EAAE,KAAK,EAAE,CAAC,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAA;IACtE,CAAC;IACD,OAAO,SAAS,CAAA;AAClB,CAAC"}

package/dist/sql-splitter.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"sql-splitter.d.ts","sourceRoot":"","sources":["../src/sql-splitter.ts"],"names":[],"mappings":"AAAA,wBAAgB,kBAAkB,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,EAAE,CA0FxD;AAED,wBAAgB,2BAA2B,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,EAAE,CAkGjE"}
+{"version":3,"file":"sql-splitter.d.ts","sourceRoot":"","sources":["../src/sql-splitter.ts"],"names":[],"mappings":"AAEA,wBAAgB,kBAAkB,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,EAAE,CAoGxD;AAED,wBAAgB,2BAA2B,CAAC,GAAG,EAAE,MAAM,GAAG,MAAM,EAAE,CAsHjE"}

package/dist/sql-splitter.js

@@ -1,3 +1,4 @@
+import { isTrivia, tokenizeSQL } from './sql-lexer.js';
 export function splitSqlStatements(sql) {
     const statements = [];
     let current = '';
@@ -19,19 +20,29 @@ export function splitSqlStatements(sql) {
         }
         if (inSingleQuote) {
             current += ch;
+            if (ch === '\\' && next) {
+                current += next;
+                i += 1;
+                continue;
+            }
             if (ch === "'" && next === "'") {
                 current += next;
                 i += 1;
                 continue;
             }
-            if (ch === "'" && sql[i - 1] !== '\\') {
+            if (ch === "'") {
                 inSingleQuote = false;
             }
             continue;
         }
         if (inDoubleQuote) {
             current += ch;
-            if (ch === '"' && sql[i - 1] !== '\\') {
+            if (ch === '\\' && next) {
+                current += next;
+                i += 1;
+                continue;
+            }
+            if (ch === '"') {
                 inDoubleQuote = false;
             }
             continue;
@@ -110,19 +121,29 @@ export function extractExecutableStatements(sql) {
         }
         if (inSingleQuote) {
             stripped += ch;
+            if (ch === '\\' && next) {
+                stripped += next;
+                i += 1;
+                continue;
+            }
             if (ch === "'" && next === "'") {
                 stripped += next;
                 i += 1;
                 continue;
             }
-            if (ch === "'" && sql[i - 1] !== '\\') {
+            if (ch === "'") {
                 inSingleQuote = false;
             }
             continue;
         }
         if (inDoubleQuote) {
             stripped += ch;
-            if (ch === '"' && sql[i - 1] !== '\\') {
+            if (ch === '\\' && next) {
+                stripped += next;
+                i += 1;
+                continue;
+            }
+            if (ch === '"') {
                 inDoubleQuote = false;
             }
             continue;
@@ -163,6 +184,32 @@ export function extractExecutableStatements(sql) {
         }
         stripped += ch;
     }
-    return splitSqlStatements(stripped);
+    const statements = splitSqlStatements(stripped);
+    // A statement made only of comments is not executable: ClickHouse rejects it
+    // as "Empty query" (for example a block comment after the last statement).
+    const executable = statements.filter(hasExecutableContent);
+    if (executable.length === statements.length)
+        return statements;
+    // The splitter does not know `#` and `//` line comments (#197). When one of
+    // them holds a `;`, the splitter cut that comment line in two; keeping the
+    // comment-only piece makes ClickHouse reject the file instead of running the
+    // commented-out text after the `;`.
+    if (tokenizeSQL(stripped).some(isLineCommentWithSemicolon))
+        return statements;
+    return executable;
+}
+// True unless every token is whitespace or a terminated comment (`/* */`,
+// nested or not, `# `, `#!`, `//`; `--` comments are already stripped). An
+// unterminated comment counts as content, so ClickHouse reports it.
+function hasExecutableContent(statement) {
+    // Only a statement that starts like a comment can be made only of comments.
+    if (!/^[-/#]/.test(statement))
+        return true;
+    // The splitter appended the `;`; a trailing line comment would swallow it.
+    const body = statement.endsWith(';') ? statement.slice(0, -1) : statement;
+    return tokenizeSQL(body).some((token) => !isTrivia(token) || !token.terminated);
+}
+function isLineCommentWithSemicolon(token) {
+    return token.kind === 'line_comment' && token.text.includes(';');
 }
 //# sourceMappingURL=sql-splitter.js.map