[ES|QL] Improve column parsing (elastic#190612)

## Summary

Fixes elastic#189913

- Adds `parts` field to `{ type: 'column' }` AST nodes. The `parts`
field contains a list of correctly parsed out column framents.
- The existing `text` and `name` fields work only if a column does not
have any escaped characters. However, I could not change/remove them, as
lots of code is currently referring them.


### Checklist

Delete any items that are not applicable to this PR.

- [x] [Unit or functional
tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html)
were updated or added to match the most common scenarios

### For maintainers

- [x] This was checked for breaking API changes and was [labeled
appropriately](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)

---------

Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com>
Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com>

packages/kbn-esql-ast/src/__tests__/ast_parser.columns.test.ts

@@ -0,0 +1,92 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0 and the Server Side Public License, v 1; you may not use this file except
+ * in compliance with, at your election, the Elastic License 2.0 or the Server
+ * Side Public License, v 1.
+ */
+
+import { getAstAndSyntaxErrors as parse } from '../ast_parser';
+
+describe('Column Identifier Expressions', () => {
+  it('can parse un-quoted identifiers', () => {
+    const text = 'ROW a, b.c';
+    const { ast } = parse(text);
+
+    expect(ast).toMatchObject([
+      {
+        type: 'command',
+        args: [
+          {
+            type: 'column',
+            parts: ['a'],
+          },
+          {
+            type: 'column',
+            parts: ['b', 'c'],
+          },
+        ],
+      },
+    ]);
+  });
+
+  it('can parse quoted identifiers', () => {
+    const text = 'ROW `a`, `b`.c, `d`.`👍`.`123``123`';
+    const { ast } = parse(text);
+
+    expect(ast).toMatchObject([
+      {
+        type: 'command',
+        args: [
+          {
+            type: 'column',
+            parts: ['a'],
+          },
+          {
+            type: 'column',
+            parts: ['b', 'c'],
+          },
+          {
+            type: 'column',
+            parts: ['d', '👍', '123`123'],
+          },
+        ],
+      },
+    ]);
+  });
+
+  it('can mix quoted and un-quoted identifiers', () => {
+    const text = 'ROW part1.part2.`part``3️⃣`';
+    const { ast } = parse(text);
+
+    expect(ast).toMatchObject([
+      {
+        type: 'command',
+        args: [
+          {
+            type: 'column',
+            parts: ['part1', 'part2', 'part`3️⃣'],
+          },
+        ],
+      },
+    ]);
+  });
+
+  it('in KEEP command', () => {
+    const text = 'FROM a | KEEP a.b';
+    const { ast } = parse(text);
+
+    expect(ast).toMatchObject([
+      {},
+      {
+        type: 'command',
+        args: [
+          {
+            type: 'column',
+            parts: ['a', 'b'],
+          },
+        ],
+      },
+    ]);
+  });
+});

packages/kbn-esql-ast/src/ast_helpers.ts

@@ -11,12 +11,14 @@
  */
 
 import { type Token, type ParserRuleContext, type TerminalNode } from 'antlr4';
-import type {
-  ArithmeticUnaryContext,
-  DecimalValueContext,
-  InlineCastContext,
-  IntegerValueContext,
-  QualifiedIntegerLiteralContext,
+import {
+  QualifiedNameContext,
+  type ArithmeticUnaryContext,
+  type DecimalValueContext,
+  type InlineCastContext,
+  type IntegerValueContext,
+  type QualifiedIntegerLiteralContext,
+  QualifiedNamePatternContext,
 } from './antlr/esql_parser';
 import { getPosition } from './ast_position_utils';
 import { DOUBLE_TICKS_REGEX, SINGLE_BACKTICK, TICKS_REGEX } from './constants';
@@ -39,6 +41,7 @@ import type {
   FunctionSubtype,
   ESQLNumericLiteral,
 } from './types';
+import { parseIdentifier } from './parser/helpers';
 
 export function nonNullable<T>(v: T): v is NonNullable<T> {
   return v != null;
@@ -360,22 +363,36 @@ export function createSource(
 }
 
 export function createColumnStar(ctx: TerminalNode): ESQLColumn {
+  const text = ctx.getText();
+
   return {
     type: 'column',
-    name: ctx.getText(),
-    text: ctx.getText(),
+    name: text,
+    parts: [text],
+    text,
     location: getPosition(ctx.symbol),
     incomplete: ctx.getText() === '',
     quoted: false,
   };
 }
 
 export function createColumn(ctx: ParserRuleContext): ESQLColumn {
+  const parts: string[] = [];
+  if (ctx instanceof QualifiedNamePatternContext) {
+    parts.push(
+      ...ctx.identifierPattern_list().map((identifier) => parseIdentifier(identifier.getText()))
+    );
+  } else if (ctx instanceof QualifiedNameContext) {
+    parts.push(...ctx.identifier_list().map((identifier) => parseIdentifier(identifier.getText())));
+  } else {
+    parts.push(sanitizeIdentifierString(ctx));
+  }
   const text = sanitizeIdentifierString(ctx);
   const hasQuotes = Boolean(getQuotedText(ctx) || isQuoted(ctx.getText()));
   return {
     type: 'column' as const,
     name: text,
+    parts,
     text: ctx.getText(),
     location: getPosition(ctx.start, ctx.stop),
     incomplete: Boolean(ctx.exception || text === ''),

packages/kbn-esql-ast/src/parser/helpers.ts

@@ -0,0 +1,37 @@
+/*
+ * Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
+ * or more contributor license agreements. Licensed under the Elastic License
+ * 2.0 and the Server Side Public License, v 1; you may not use this file except
+ * in compliance with, at your election, the Elastic License 2.0 or the Server
+ * Side Public License, v 1.
+ */
+
+export const isQuotedIdentifier = (text: string): boolean => {
+  const firstChar = text[0];
+  const lastChar = text[text.length - 1];
+
+  return firstChar === '`' && lastChar === '`';
+};
+
+export const parseIdentifier = (text: string): string => {
+  const isQuoted = isQuotedIdentifier(text);
+
+  if (!isQuoted) {
+    return text;
+  }
+
+  return text.slice(1, -1).replace(/``/g, '`');
+};
+
+export const regexUnquotedIdentifierPattern = /^([a-z\*_\@]{1})[a-z0-9_\*]*$/i;
+
+export const formatIdentifier = (text: string): string => {
+  if (regexUnquotedIdentifierPattern.test(text)) {
+    return text;
+  }
+
+  return `\`${text.replace(/`/g, '``')}\``;
+};
+
+export const formatIdentifierParts = (parts: string[]): string =>
+  parts.map(formatIdentifier).join('.');

packages/kbn-esql-ast/src/types.ts

@@ -179,6 +179,22 @@ export interface ESQLSource extends ESQLAstBaseItem {
 
 export interface ESQLColumn extends ESQLAstBaseItem {
   type: 'column';
+
+  /**
+   * An identifier can be composed of multiple parts, e.g: part1.part2.`part``3️⃣`.
+   * This property contains the parsed unquoted parts of the identifier.
+   * For example: `['part1', 'part2', 'part`3️⃣']`.
+   */
+  parts: string[];
+
+  /**
+   * @deprecated
+   *
+   * An identifier can be composed of multiple parts, e.g: part1.part2.`part3️⃣`
+   *
+   * Each part can be quoted or not quoted independently. A single `quoted`
+   * property is not enough to represent the identifier. Use `parts` instead.
+   */
   quoted: boolean;
 }