WithStatement::build() - Code Metrics - Inspection of "Merge branch '5.9.x'" - phpmyadmin/sql-parser - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — master ( eccca0...6f769c )

by Maurício

created 2023-09-16 14:39 UTC

WithStatement::build() A

↳ Parent: WithStatement

Complexity

Conditions	5
Paths	6

Size

Total Lines	18
Code Lines	9

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	10
CRAP Score	5

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
cc	5
eloc	9
nc	6
nop	0
dl	0
loc	18
ccs	10
cts	10
cp	1
crap	5
rs	9.6111
c	1
b	0
f	0

<?php

declare(strict_types=1);

namespace PhpMyAdmin\SqlParser\Statements;

use PhpMyAdmin\SqlParser\Components\Array2d;
use PhpMyAdmin\SqlParser\Components\OptionsArray;
use PhpMyAdmin\SqlParser\Components\WithKeyword;
use PhpMyAdmin\SqlParser\Exceptions\ParserException;
use PhpMyAdmin\SqlParser\Parser;
use PhpMyAdmin\SqlParser\Statement;
use PhpMyAdmin\SqlParser\Token;
use PhpMyAdmin\SqlParser\TokensList;
use PhpMyAdmin\SqlParser\Translator;

use function array_slice;
use function preg_match;

/**
 * `WITH` statement.

 *  WITH [RECURSIVE] query_name [ (column_name [,...]) ] AS (SELECT ...) [, ...]
 */
final class WithStatement extends Statement
{
    /**
     * Options for `WITH` statements and their slot ID.
     *
     * @var array<string, int|array<int, int|string>>
     * @psalm-var array<string, (positive-int|array{positive-int, ('var'|'var='|'expr'|'expr=')})>
     */
    public static $statementOptions = ['RECURSIVE' => 1];

    /**
     * The clauses of this statement, in order.
     *
     * @see Statement::$clauses
     *
     * @var array<string, array<int, int|string>>
     * @psalm-var array<string, array{non-empty-string, (1|2|3)}>
     */
    public static $clauses = [
        'WITH' => [
            'WITH',
            2,
        ],
        // Used for options.
        '_OPTIONS' => [
            '_OPTIONS',
            1,
        ],
        'AS' => [
            'AS',
            2,
        ],
    ];

    /** @var WithKeyword[] */
    public $withers = [];

    /**
     * holds the CTE parser.
     *
     * @var Parser|null
     */
    public $cteStatementParser;

    /**
     * @param Parser     $parser the instance that requests parsing
     * @param TokensList $list   the list of tokens to be parsed
     */
    public function parse(Parser $parser, TokensList $list): void
    {
        /**
         * The state of the parser.
         *
         * Below are the states of the parser.
         *
         *      0 ---------------- [ name ] -----------------> 1
         *
         *      1 ------------------ [ ( ] ------------------> 2
         *
         *      2 ------------------ [ AS ] -----------------> 3
         *
         *      3 ------------------ [ ( ] ------------------> 4
         *
         *      4 ------------------ [ , ] ------------------> 1
         *
         *      4 ----- [ SELECT/UPDATE/DELETE/INSERT ] -----> 5
         *
         * @var int
         */
        $state = 0;
        $wither = null;

        ++$list->idx; // Skipping `WITH`.

        // parse any options if provided
        $this->options = OptionsArray::parse($parser, $list, static::$statementOptions);
        ++$list->idx;

        for (; $list->idx < $list->count; ++$list->idx) {
            /**
             * Token parsed at this moment.
             */
            $token = $list->tokens[$list->idx];

            // Skipping whitespaces and comments.
            if ($token->type === Token::TYPE_WHITESPACE || $token->type === Token::TYPE_COMMENT) {
                continue;
            }

            if ($state === 0) {
                if ($token->type !== Token::TYPE_NONE || ! preg_match('/^[a-zA-Z0-9_$]+$/', $token->token)) {
                    $parser->error('The name of the CTE was expected.', $token);
                    break;
                }

                $wither = $token->value;
                $this->withers[$wither] = new WithKeyword($wither);
                $state = 1;
            } elseif ($state === 1) {
                if ($token->type === Token::TYPE_OPERATOR && $token->value === '(') {
                    $columns = Array2d::parse($parser, $list);
                    if ($parser->errors !== []) {
                        break;
                    }

                    $this->withers[$wither]->columns = $columns;
                    $state = 2;
                } elseif ($token->type === Token::TYPE_KEYWORD && $token->keyword === 'AS') {
                    $state = 3;
                } else {
                    $parser->error('Unexpected token.', $token);
                    break;
                }
            } elseif ($state === 2) {
                if (! ($token->type === Token::TYPE_KEYWORD && $token->keyword === 'AS')) {
                    $parser->error('AS keyword was expected.', $token);
                    break;
                }

                $state = 3;
            } elseif ($state === 3) {
                $idxBeforeGetNext = $list->idx;

                $list->idx++; // Ignore the current token
                $nextKeyword = $list->getNext();

                if (! ($token->value === '(' && ($nextKeyword && $nextKeyword->value === 'SELECT'))) {
                    $parser->error('Subquery of the CTE was expected.', $token);
                    $list->idx = $idxBeforeGetNext;
                    break;
                }

                // Restore the index
                $list->idx = $idxBeforeGetNext;

                ++$list->idx;
                $subList = $this->getSubTokenList($list);
                if ($subList instanceof ParserException) {
                    $parser->errors[] = $subList;
                    break;
                }

                $subParser = new Parser($subList);

                if ($subParser->errors !== []) {
                    foreach ($subParser->errors as $error) {
                        $parser->errors[] = $error;
                    }

                    break;
                }

                $this->withers[$wither]->statement = $subParser;

                $state = 4;
            } elseif ($state === 4) {
                if ($token->value === ',') {
                    // There's another WITH expression to parse, go back to state=0
                    $state = 0;
                    continue;
                }

                if (
                    $token->type === Token::TYPE_KEYWORD && (
                    $token->value === 'SELECT'
                    || $token->value === 'INSERT'
                    || $token->value === 'UPDATE'
                    || $token->value === 'DELETE'
                    )
                ) {
                    $state = 5;
                    --$list->idx;
                    continue;
                }

                $parser->error('An expression was expected.', $token);
                break;
            } elseif ($state === 5) {
                /**
                 * We need to parse all of the remaining tokens becuase mostly, they are only the CTE expression
                 * which's mostly is SELECT, or INSERT, UPDATE, or delete statement.
                 * e.g: INSERT .. ( SELECT 1 ) SELECT col1 FROM cte ON DUPLICATE KEY UPDATE col_name = 3.
                 * The issue is that, `ON DUPLICATE KEY UPDATE col_name = 3` is related to the main INSERT query
                 * not the cte expression (SELECT col1 FROM cte) we need to determine the end of the expression
                 * to parse `ON DUPLICATE KEY UPDATE` from the InsertStatement parser instead.
                 */

                // Index of the last parsed token by default would be the last token in the $list, because we're
                // assuming that all remaining tokens at state 4, are related to the expression.
                $idxOfLastParsedToken = $list->count - 1;
                // Index before search to be able to restore the index.
                $idxBeforeSearch = $list->idx;
                // Length of expression tokens is null by default, in order for the $subList to start
                // from $list->idx to the end of the $list.
                $lengthOfExpressionTokens = null;

                if ($list->getNextOfTypeAndValue(Token::TYPE_KEYWORD, 'ON')) {
                    // (-1) because getNextOfTypeAndValue returned ON and increased the index.
                    $idxOfOn = $list->idx - 1;
                    // We want to make sure that it's `ON DUPLICATE KEY UPDATE`
                    $dubplicateToken = $list->getNext();
                    $keyToken = $list->getNext();
                    $updateToken = $list->getNext();
                    if (
                        $dubplicateToken && $dubplicateToken->keyword === 'DUPLICATE'
                        && ($keyToken && $keyToken->keyword === 'KEY')
                        && ($updateToken && $updateToken->keyword === 'UPDATE')
                    ) {
                        // Index of the last parsed token will be the token before the ON Keyword
                        $idxOfLastParsedToken = $idxOfOn - 1;
                        // The length of the expression tokens would be the difference
                        // between the first unrelated token `ON` and the idx
                        // before skipping the CTE tokens.
                        $lengthOfExpressionTokens = $idxOfOn - $idxBeforeSearch;
                    }
                }

                // Restore the index
                $list->idx = $idxBeforeSearch;

                $subList = new TokensList(array_slice($list->tokens, $list->idx, $lengthOfExpressionTokens));
                $subParser = new Parser($subList);
                if ($subParser->errors !== []) {
                    foreach ($subParser->errors as $error) {
                        $parser->errors[] = $error;
                    }

                    break;
                }

                $this->cteStatementParser = $subParser;

                $list->idx = $idxOfLastParsedToken;
                break;
            }
        }

        // 5 is the only valid end state
        if ($state !== 5) {

             /**
             * Token parsed at this moment.
             */
            $token = $list->tokens[$list->idx];

            $parser->error('Unexpected end of the WITH CTE.', $token);
        }

        --$list->idx;
    }

    public function build(): string
    {
        $str = 'WITH ';

        foreach ($this->withers as $wither) {
            $str .= $str === 'WITH ' ? '' : ', ';
            $str .= WithKeyword::build($wither);
        }

        $str .= ' ';

        if ($this->cteStatementParser) {
            foreach ($this->cteStatementParser->statements as $statement) {
                    $str .= $statement->build();
            }
        }

        return $str;
    }

    /**
     * Get tokens within the WITH expression to use them in another parser
     */
    private function getSubTokenList(TokensList $list): ParserException|TokensList
    {
        $idx = $list->idx;
        $token = $list->tokens[$list->idx];
        $openParenthesis = 0;

        while ($list->idx < $list->count) {
            if ($token->value === '(') {
                ++$openParenthesis;
            } elseif ($token->value === ')') {
                if (--$openParenthesis === -1) {
                    break;
                }
            }

            ++$list->idx;
            if (! isset($list->tokens[$list->idx])) {
                break;
            }

            $token = $list->tokens[$list->idx];
        }

        // performance improvement: return the error to avoid a try/catch in the loop
        if ($list->idx === $list->count) {
            --$list->idx;

            return new ParserException(
                Translator::gettext('A closing bracket was expected.'),
                $token
            );
        }

        $length = $list->idx - $idx;

        return new TokensList(array_slice($list->tokens, $idx, $length), $length);
    }
}


1		<?php
2
3		declare(strict_types=1);
4
5		namespace PhpMyAdmin\SqlParser\Statements;
6
7		use PhpMyAdmin\SqlParser\Components\Array2d;
8		use PhpMyAdmin\SqlParser\Components\OptionsArray;
9		use PhpMyAdmin\SqlParser\Components\WithKeyword;
10		use PhpMyAdmin\SqlParser\Exceptions\ParserException;
11		use PhpMyAdmin\SqlParser\Parser;
12		use PhpMyAdmin\SqlParser\Statement;
13		use PhpMyAdmin\SqlParser\Token;
14		use PhpMyAdmin\SqlParser\TokensList;
15		use PhpMyAdmin\SqlParser\Translator;
16
17		use function array_slice;
18		use function preg_match;
19
20		/**
21		* `WITH` statement.
22
23		* WITH [RECURSIVE] query_name [ (column_name [,...]) ] AS (SELECT ...) [, ...]
24		*/
25		final class WithStatement extends Statement
26		{
27		/**
28		* Options for `WITH` statements and their slot ID.
29		*
30		* @var array<string, int\|array<int, int\|string>>
31		* @psalm-var array<string, (positive-int\|array{positive-int, ('var'\|'var='\|'expr'\|'expr=')})>
32		*/
33		public static $statementOptions = ['RECURSIVE' => 1];
34
35		/**
36		* The clauses of this statement, in order.
37		*
38		* @see Statement::$clauses
39		*
40		* @var array<string, array<int, int\|string>>
41		* @psalm-var array<string, array{non-empty-string, (1\|2\|3)}>
42		*/
43		public static $clauses = [
44		'WITH' => [
45		'WITH',
46		2,
47		],
48		// Used for options.
49		'_OPTIONS' => [
50		'_OPTIONS',
51		1,
52		],
53		'AS' => [
54		'AS',
55		2,
56		],
57		];
58
59		/** @var WithKeyword[] */
60		public $withers = [];
61
62		/**
63		* holds the CTE parser.
64		*
65		* @var Parser\|null
66		*/
67		public $cteStatementParser;
68
69		/**
70		* @param Parser $parser the instance that requests parsing
71		* @param TokensList $list the list of tokens to be parsed
72		*/
73	52	public function parse(Parser $parser, TokensList $list): void
74		{
75		/**
76		* The state of the parser.
77		*
78		* Below are the states of the parser.
79		*
80		* 0 ---------------- [ name ] -----------------> 1
81		*
82		* 1 ------------------ [ ( ] ------------------> 2
83		*
84		* 2 ------------------ [ AS ] -----------------> 3
85		*
86		* 3 ------------------ [ ( ] ------------------> 4
87		*
88		* 4 ------------------ [ , ] ------------------> 1
89		*
90		* 4 ----- [ SELECT/UPDATE/DELETE/INSERT ] -----> 5
91		*
92		* @var int
93		*/
94	52	$state = 0;
95	52	$wither = null;
96
97	52	++$list->idx; // Skipping `WITH`.
98
99		// parse any options if provided
100	52	$this->options = OptionsArray::parse($parser, $list, static::$statementOptions);
101	52	++$list->idx;
102
103	52	for (; $list->idx < $list->count; ++$list->idx) {
104		/**
105		* Token parsed at this moment.
106		*/
107	52	$token = $list->tokens[$list->idx];
108
109		// Skipping whitespaces and comments.
110	52	if ($token->type === Token::TYPE_WHITESPACE \|\| $token->type === Token::TYPE_COMMENT) {
111	46	continue;
112		}
113
114	52	if ($state === 0) {
115	52	if ($token->type !== Token::TYPE_NONE \|\| ! preg_match('/^[a-zA-Z0-9_$]+$/', $token->token)) {
116	6	$parser->error('The name of the CTE was expected.', $token);
117	6	break;
118		}
119
120	48	$wither = $token->value;
121	48	$this->withers[$wither] = new WithKeyword($wither);
122	48	$state = 1;
123	48	} elseif ($state === 1) {
124	48	if ($token->type === Token::TYPE_OPERATOR && $token->value === '(') {
125	40	$columns = Array2d::parse($parser, $list);
126	40	if ($parser->errors !== []) {
127	2	break;
128		}
129
130	38	$this->withers[$wither]->columns = $columns;
131	38	$state = 2;
132	12	} elseif ($token->type === Token::TYPE_KEYWORD && $token->keyword === 'AS') {
133	10	$state = 3;
134		} else {
135	2	$parser->error('Unexpected token.', $token);
136	24	break;
137		}
138	44	} elseif ($state === 2) {
139	38	if (! ($token->type === Token::TYPE_KEYWORD && $token->keyword === 'AS')) {
140	2	$parser->error('AS keyword was expected.', $token);
141	2	break;
142		}
143
144	36	$state = 3;
145	42	} elseif ($state === 3) {
146	42	$idxBeforeGetNext = $list->idx;
147
148	42	$list->idx++; // Ignore the current token
149	42	$nextKeyword = $list->getNext();
150
151	42	if (! ($token->value === '(' && ($nextKeyword && $nextKeyword->value === 'SELECT'))) {
152	4	$parser->error('Subquery of the CTE was expected.', $token);
153	4	$list->idx = $idxBeforeGetNext;
154	4	break;
155		}
156
157		// Restore the index
158	38	$list->idx = $idxBeforeGetNext;
159
160	38	++$list->idx;
161	38	$subList = $this->getSubTokenList($list);
162	38	if ($subList instanceof ParserException) {
163	2	$parser->errors[] = $subList;
164	2	break;
165		}
166
167	36	$subParser = new Parser($subList);
168
169	36	if ($subParser->errors !== []) {
170	2	foreach ($subParser->errors as $error) {
171	2	$parser->errors[] = $error;
172		}
173
174	2	break;
175		}
176
177	34	$this->withers[$wither]->statement = $subParser;
178
179	34	$state = 4;
180	34	} elseif ($state === 4) {
181	34	if ($token->value === ',') {
182		// There's another WITH expression to parse, go back to state=0
183	10	$state = 0;
184	10	continue;
185		}
186
187		if (
188	32	$token->type === Token::TYPE_KEYWORD && (
189	32	$token->value === 'SELECT'
190	32	\|\| $token->value === 'INSERT'
191	32	\|\| $token->value === 'UPDATE'
192	32	\|\| $token->value === 'DELETE'
193		)
194		) {
195	26	$state = 5;
196	26	--$list->idx;
197	26	continue;
198		}
199
200	6	$parser->error('An expression was expected.', $token);
201	6	break;
202	26	} elseif ($state === 5) {
203		/**
204		* We need to parse all of the remaining tokens becuase mostly, they are only the CTE expression
205		* which's mostly is SELECT, or INSERT, UPDATE, or delete statement.
206		* e.g: INSERT .. ( SELECT 1 ) SELECT col1 FROM cte ON DUPLICATE KEY UPDATE col_name = 3.
207		* The issue is that, `ON DUPLICATE KEY UPDATE col_name = 3` is related to the main INSERT query
208		* not the cte expression (SELECT col1 FROM cte) we need to determine the end of the expression
209		* to parse `ON DUPLICATE KEY UPDATE` from the InsertStatement parser instead.
210		*/
211
212		// Index of the last parsed token by default would be the last token in the $list, because we're
213		// assuming that all remaining tokens at state 4, are related to the expression.
214	26	$idxOfLastParsedToken = $list->count - 1;
215		// Index before search to be able to restore the index.
216	26	$idxBeforeSearch = $list->idx;
217		// Length of expression tokens is null by default, in order for the $subList to start
218		// from $list->idx to the end of the $list.
219	26	$lengthOfExpressionTokens = null;
220
221	26	if ($list->getNextOfTypeAndValue(Token::TYPE_KEYWORD, 'ON')) {
222		// (-1) because getNextOfTypeAndValue returned ON and increased the index.
223	4	$idxOfOn = $list->idx - 1;
224		// We want to make sure that it's `ON DUPLICATE KEY UPDATE`
225	4	$dubplicateToken = $list->getNext();
226	4	$keyToken = $list->getNext();
227	4	$updateToken = $list->getNext();
228		if (
229	4	$dubplicateToken && $dubplicateToken->keyword === 'DUPLICATE'
230	4	&& ($keyToken && $keyToken->keyword === 'KEY')
231	4	&& ($updateToken && $updateToken->keyword === 'UPDATE')
232		) {
233		// Index of the last parsed token will be the token before the ON Keyword
234	2	$idxOfLastParsedToken = $idxOfOn - 1;
235		// The length of the expression tokens would be the difference
236		// between the first unrelated token `ON` and the idx
237		// before skipping the CTE tokens.
238	2	$lengthOfExpressionTokens = $idxOfOn - $idxBeforeSearch;
239		}
240		}
241
242		// Restore the index
243	26	$list->idx = $idxBeforeSearch;
244
245	26	$subList = new TokensList(array_slice($list->tokens, $list->idx, $lengthOfExpressionTokens));
246	26	$subParser = new Parser($subList);
247	26	if ($subParser->errors !== []) {
248	2	foreach ($subParser->errors as $error) {
249	2	$parser->errors[] = $error;
250		}
251
252	2	break;
253		}
254
255	24	$this->cteStatementParser = $subParser;
256
257	24	$list->idx = $idxOfLastParsedToken;
258	24	break;
259		}
260		}
261
262		// 5 is the only valid end state
263	52	if ($state !== 5) {
		0 ignored issues – show introduced 2021-12-06 01:53 UTC by Report Bug Copy Issue Report The condition `$state !== 5` is always `true`. Loading history...
264		/**
265		* Token parsed at this moment.
266		*/
267	26	$token = $list->tokens[$list->idx];
268
269	26	$parser->error('Unexpected end of the WITH CTE.', $token);
270		}
271
272	52	--$list->idx;
273		}
274
275	10	public function build(): string
276		{
277	10	$str = 'WITH ';
278
279	10	foreach ($this->withers as $wither) {
280	10	$str .= $str === 'WITH ' ? '' : ', ';
281	10	$str .= WithKeyword::build($wither);
282		}
283
284	10	$str .= ' ';
285
286	10	if ($this->cteStatementParser) {
287	10	foreach ($this->cteStatementParser->statements as $statement) {
288	10	$str .= $statement->build();
289		}
290		}
291
292	10	return $str;
293		}
294
295		/**
296		* Get tokens within the WITH expression to use them in another parser
297		*/
298	38	private function getSubTokenList(TokensList $list): ParserException\|TokensList
299		{
300	38	$idx = $list->idx;
301	38	$token = $list->tokens[$list->idx];
302	38	$openParenthesis = 0;
303
304	38	while ($list->idx < $list->count) {
305	38	if ($token->value === '(') {
306	4	++$openParenthesis;
307	38	} elseif ($token->value === ')') {
308	38	if (--$openParenthesis === -1) {
309	36	break;
310		}
311		}
312
313	38	++$list->idx;
314	38	if (! isset($list->tokens[$list->idx])) {
315	2	break;
316		}
317
318	38	$token = $list->tokens[$list->idx];
319		}
320
321		// performance improvement: return the error to avoid a try/catch in the loop
322	38	if ($list->idx === $list->count) {
323	2	--$list->idx;
324
325	2	return new ParserException(
326	2	Translator::gettext('A closing bracket was expected.'),
327	2	$token
328	2	);
329		}
330
331	36	$length = $list->idx - $idx;
332
333	36	return new TokensList(array_slice($list->tokens, $idx, $length), $length);
334		}
335		}
336

phpmyadmin / sql-parser

Push — master ( eccca0...6f769c )

WithStatement::build() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like