PostgresArrayToPHPArrayTransformer::parsePostgresArrayManually() - Code Metrics - Inspection of "fix(#351): correct round-trip handling of special..." - martin-georgiev/postgresql-for-doctrine - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Pull Request — main (#391)

by Martin

created 2025-06-14 23:40 UTC

parsePostgresArrayManually() B

↳ Parent: PostgresArrayToPHPArrayTransformer

Complexity

Conditions	10
Paths	13

Size

Total Lines	49
Code Lines	27

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	26
CRAP Score	10.005

Importance

Changes	1
Bugs	1	Features	0

Metric	Value
cc	10
eloc	27
c	1
b	1
f	0
nc	13
nop	1
dl	0
loc	49
ccs	26
cts	27
cp	0.963
crap	10.005
rs	7.6666

How to fix Complexity

<?php

declare(strict_types=1);

namespace MartinGeorgiev\Utils;

use MartinGeorgiev\Utils\Exception\InvalidArrayFormatException;

/**
 * Handles transformation from PostgreSQL text arrays to PHP values.
 *
 * @since 3.0
 *
 * @author Martin Georgiev <[email protected]>
 */
class PostgresArrayToPHPArrayTransformer
{
    private const POSTGRESQL_EMPTY_ARRAY = '{}';

    private const POSTGRESQL_NULL_VALUE = 'null';

    /**
     * Transforms a PostgreSQL text array to a PHP array.
     * This method supports only single-dimensional text arrays and
     * relies on the default escaping strategy in PostgreSQL (double quotes).
     *
     * @throws InvalidArrayFormatException when the input is a multi-dimensional array or has an invalid format
     */
    public static function transformPostgresArrayToPHPArray(string $postgresArray): array
    {
        $trimmed = \trim($postgresArray);

        if ($trimmed === '' || \strtolower($trimmed) === self::POSTGRESQL_NULL_VALUE) {
            return [];
        }

        if (\str_contains($trimmed, '},{') || \str_starts_with($trimmed, '{{')) {
            throw InvalidArrayFormatException::multiDimensionalArrayNotSupported();
        }

        if ($trimmed === self::POSTGRESQL_EMPTY_ARRAY) {
            return [];
        }

        // Check for malformed nesting - this is a more specific check than the one above
        // But we need to exclude cases where curly braces are part of quoted strings
        $content = \trim($trimmed, self::POSTGRESQL_EMPTY_ARRAY);
        $inQuotes = false;
        $escaping = false;

        for ($i = 0, $len = \strlen($content); $i < $len; $i++) {
            $char = $content[$i];

            if ($escaping) {
                $escaping = false;

                continue;
            }

            if ($char === '\\' && $inQuotes) {
                $escaping = true;

                continue;
            }

            if ($char === '"') {
                $inQuotes = !$inQuotes;

            } elseif (($char === '{' || $char === '}') && !$inQuotes) {
                throw InvalidArrayFormatException::invalidFormat('Malformed array nesting detected');
            }
        }

        // Check for unclosed quotes
        if ($inQuotes) {

            throw InvalidArrayFormatException::invalidFormat('Unclosed quotes in array');
        }

        // First try with json_decode for properly quoted values
        $jsonArray = '['.\trim($trimmed, self::POSTGRESQL_EMPTY_ARRAY).']';

        /** @var array<int, mixed>|null $decoded */
        $decoded = \json_decode($jsonArray, true, 512, JSON_BIGINT_AS_STRING);

        // If json_decode fails, try manual parsing for unquoted strings
        if ($decoded === null && \json_last_error() !== JSON_ERROR_NONE) {
            return self::parsePostgresArrayManually($content);
        }

        return (array) $decoded;
    }

    private static function parsePostgresArrayManually(string $content): array
    {
        if ($content === '') {
            return [];
        }

        // Parse the array manually, handling quoted and unquoted values
        $result = [];
        $inQuotes = false;
        $currentValue = '';
        $escaping = false;

        for ($i = 0, $len = \strlen($content); $i < $len; $i++) {
            $char = $content[$i];

            // Handle escaping within quotes
            if ($escaping) {
                $currentValue .= $char;
                $escaping = false;

                continue;
            }

            if ($char === '\\' && $inQuotes) {
                $escaping = true;
                $currentValue .= $char;

                continue;
            }

            if ($char === '"') {
                $inQuotes = !$inQuotes;

                // For quoted values, we include the quotes for later processing
                $currentValue .= $char;
            } elseif ($char === ',' && !$inQuotes) {
                // End of value
                $result[] = self::processPostgresValue($currentValue);
                $currentValue = '';
            } else {
                $currentValue .= $char;
            }
        }

        // Add the last value
        if ($currentValue !== '') {
            $result[] = self::processPostgresValue($currentValue);
        }

        return $result;
    }

    /**
     * Process a single value from a PostgreSQL array.
     */
    private static function processPostgresValue(string $value): mixed
    {
        $value = \trim($value);

        if (self::isNullValue($value)) {
            return null;
        }

        if (self::isBooleanValue($value)) {
            return self::processBooleanValue($value);
        }

        if (self::isQuotedString($value)) {
            return self::processQuotedString($value);
        }

        if (self::isNumericValue($value)) {
            return self::processNumericValue($value);
        }

        // For unquoted strings, return as is
        return $value;
    }

    private static function isNullValue(string $value): bool
    {
        return $value === 'NULL' || $value === 'null';
    }

    private static function isBooleanValue(string $value): bool
    {
        return \in_array($value, ['true', 't', 'false', 'f'], true);
    }

    private static function processBooleanValue(string $value): bool
    {
        return $value === 'true' || $value === 't';
    }

    private static function isQuotedString(string $value): bool
    {
        return \strlen($value) >= 2 && $value[0] === '"' && $value[\strlen($value) - 1] === '"';
    }

    private static function processQuotedString(string $value): string
    {
        // Remove the quotes and unescape the string
        $unquoted = \substr($value, 1, -1);

        return self::unescapeString($unquoted);
    }

    private static function isNumericValue(string $value): bool
    {
        return \is_numeric($value);
    }

    private static function processNumericValue(string $value): float|int
    {
        // Convert to int or float as appropriate
        if (\str_contains($value, '.') || \stripos($value, 'e') !== false) {
            return (float) $value;
        }

        return (int) $value;
    }

    private static function unescapeString(string $value): string
    {
        /**
         * PostgreSQL array escaping rules:
         * \\ -> \ (escaped backslash becomes literal backslash)
         * \" -> " (escaped quote becomes literal quote)
         * Everything else remains as-is
         */
        $result = '';
        $length = \strlen($value);
        $position = 0;

        while ($position < $length) {
            if ($value[$position] === '\\' && $position + 1 < $length) {
                $nextChar = $value[$position + 1];

                if ($nextChar === '\\') {
                    // \\ -> \
                    $result .= '\\';
                    $position += 2;
                } elseif ($nextChar === '"') {
                    // \" -> "
                    $result .= '"';
                    $position += 2;
                } else {
                    // \ followed by anything else - keep the backslash
                    $result .= '\\';
                    $position++;
                }
            } else {
                $result .= $value[$position];
                $position++;
            }
        }

        return $result;
    }
}


1		<?php
2
3		declare(strict_types=1);
4
5		namespace MartinGeorgiev\Utils;
6
7		use MartinGeorgiev\Utils\Exception\InvalidArrayFormatException;
8
9		/**
10		* Handles transformation from PostgreSQL text arrays to PHP values.
11		*
12		* @since 3.0
13		*
14		* @author Martin Georgiev <[email protected]>
15		*/
16		class PostgresArrayToPHPArrayTransformer
17		{
18		private const POSTGRESQL_EMPTY_ARRAY = '{}';
19
20		private const POSTGRESQL_NULL_VALUE = 'null';
21
22		/**
23		* Transforms a PostgreSQL text array to a PHP array.
24		* This method supports only single-dimensional text arrays and
25		* relies on the default escaping strategy in PostgreSQL (double quotes).
26		*
27		* @throws InvalidArrayFormatException when the input is a multi-dimensional array or has an invalid format
28		*/
29	42	public static function transformPostgresArrayToPHPArray(string $postgresArray): array
30		{
31	42	$trimmed = \trim($postgresArray);
32
33	42	if ($trimmed === '' \|\| \strtolower($trimmed) === self::POSTGRESQL_NULL_VALUE) {
34	3	return [];
35		}
36
37	39	if (\str_contains($trimmed, '},{') \|\| \str_starts_with($trimmed, '{{')) {
38	1	throw InvalidArrayFormatException::multiDimensionalArrayNotSupported();
39		}
40
41	38	if ($trimmed === self::POSTGRESQL_EMPTY_ARRAY) {
42	1	return [];
43		}
44
45		// Check for malformed nesting - this is a more specific check than the one above
46		// But we need to exclude cases where curly braces are part of quoted strings
47	37	$content = \trim($trimmed, self::POSTGRESQL_EMPTY_ARRAY);
48	37	$inQuotes = false;
49	37	$escaping = false;
50
51	37	for ($i = 0, $len = \strlen($content); $i < $len; $i++) {
52	37	$char = $content[$i];
53
54	37	if ($escaping) {
55	15	$escaping = false;
56
57	15	continue;
58		}
59
60	37	if ($char === '\\' && $inQuotes) {
61	15	$escaping = true;
62
63	15	continue;
64		}
65
66	37	if ($char === '"') {
67	27	$inQuotes = !$inQuotes;
		0 ignored issues – show introduced 2025-04-03 00:08 UTC by Report Bug Copy Issue Report The condition `$inQuotes` is always `false`. Loading history...
68	35	} elseif (($char === '{' \|\| $char === '}') && !$inQuotes) {
69	1	throw InvalidArrayFormatException::invalidFormat('Malformed array nesting detected');
70		}
71		}
72
73		// Check for unclosed quotes
74	36	if ($inQuotes) {
		0 ignored issues – show introduced 2025-04-03 00:08 UTC by Report Bug Copy Issue Report The condition `$inQuotes` is always `false`. Loading history...
75	2	throw InvalidArrayFormatException::invalidFormat('Unclosed quotes in array');
76		}
77
78		// First try with json_decode for properly quoted values
79	34	$jsonArray = '['.\trim($trimmed, self::POSTGRESQL_EMPTY_ARRAY).']';
80
81		/** @var array<int, mixed>\|null $decoded */
82	34	$decoded = \json_decode($jsonArray, true, 512, JSON_BIGINT_AS_STRING);
83
84		// If json_decode fails, try manual parsing for unquoted strings
85	34	if ($decoded === null && \json_last_error() !== JSON_ERROR_NONE) {
86	11	return self::parsePostgresArrayManually($content);
87		}
88
89	23	return (array) $decoded;
90		}
91
92	11	private static function parsePostgresArrayManually(string $content): array
93		{
94	11	if ($content === '') {
95		return [];
96		}
97
98		// Parse the array manually, handling quoted and unquoted values
99	11	$result = [];
100	11	$inQuotes = false;
101	11	$currentValue = '';
102	11	$escaping = false;
103
104	11	for ($i = 0, $len = \strlen($content); $i < $len; $i++) {
105	11	$char = $content[$i];
106
107		// Handle escaping within quotes
108	11	if ($escaping) {
109	3	$currentValue .= $char;
110	3	$escaping = false;
111
112	3	continue;
113		}
114
115	11	if ($char === '\\' && $inQuotes) {
116	3	$escaping = true;
117	3	$currentValue .= $char;
118
119	3	continue;
120		}
121
122	11	if ($char === '"') {
123	5	$inQuotes = !$inQuotes;
		0 ignored issues – show introduced 2025-04-03 00:08 UTC by Report Bug Copy Issue Report The condition `$inQuotes` is always `false`. Loading history...
124		// For quoted values, we include the quotes for later processing
125	5	$currentValue .= $char;
126	11	} elseif ($char === ',' && !$inQuotes) {
127		// End of value
128	9	$result[] = self::processPostgresValue($currentValue);
129	9	$currentValue = '';
130		} else {
131	11	$currentValue .= $char;
132		}
133		}
134
135		// Add the last value
136	11	if ($currentValue !== '') {
137	10	$result[] = self::processPostgresValue($currentValue);
138		}
139
140	11	return $result;
141		}
142
143		/**
144		* Process a single value from a PostgreSQL array.
145		*/
146	11	private static function processPostgresValue(string $value): mixed
147		{
148	11	$value = \trim($value);
149
150	11	if (self::isNullValue($value)) {
151	1	return null;
152		}
153
154	11	if (self::isBooleanValue($value)) {
155	1	return self::processBooleanValue($value);
156		}
157
158	10	if (self::isQuotedString($value)) {
159	5	return self::processQuotedString($value);
160		}
161
162	7	if (self::isNumericValue($value)) {
163		return self::processNumericValue($value);
164		}
165
166		// For unquoted strings, return as is
167	7	return $value;
168		}
169
170	11	private static function isNullValue(string $value): bool
171		{
172	11	return $value === 'NULL' \|\| $value === 'null';
173		}
174
175	11	private static function isBooleanValue(string $value): bool
176		{
177	11	return \in_array($value, ['true', 't', 'false', 'f'], true);
178		}
179
180	1	private static function processBooleanValue(string $value): bool
181		{
182	1	return $value === 'true' \|\| $value === 't';
183		}
184
185	10	private static function isQuotedString(string $value): bool
186		{
187	10	return \strlen($value) >= 2 && $value[0] === '"' && $value[\strlen($value) - 1] === '"';
188		}
189
190	5	private static function processQuotedString(string $value): string
191		{
192		// Remove the quotes and unescape the string
193	5	$unquoted = \substr($value, 1, -1);
194
195	5	return self::unescapeString($unquoted);
196		}
197
198	7	private static function isNumericValue(string $value): bool
199		{
200	7	return \is_numeric($value);
201		}
202
203		private static function processNumericValue(string $value): float\|int
204		{
205		// Convert to int or float as appropriate
206		if (\str_contains($value, '.') \|\| \stripos($value, 'e') !== false) {
207		return (float) $value;
208		}
209
210		return (int) $value;
211		}
212
213	5	private static function unescapeString(string $value): string
214		{
215		/**
216		* PostgreSQL array escaping rules:
217		* \\ -> \ (escaped backslash becomes literal backslash)
218		* \" -> " (escaped quote becomes literal quote)
219		* Everything else remains as-is
220		*/
221	5	$result = '';
222	5	$length = \strlen($value);
223	5	$position = 0;
224
225	5	while ($position < $length) {
226	5	if ($value[$position] === '\\' && $position + 1 < $length) {
227	3	$nextChar = $value[$position + 1];
228
229	3	if ($nextChar === '\\') {
230		// \\ -> \
231	1	$result .= '\\';
232	1	$position += 2;
233	2	} elseif ($nextChar === '"') {
234		// \" -> "
235	1	$result .= '"';
236	1	$position += 2;
237		} else {
238		// \ followed by anything else - keep the backslash
239	1	$result .= '\\';
240	1	$position++;
241		}
242		} else {
243	5	$result .= $value[$position];
244	5	$position++;
245		}
246		}
247
248	5	return $result;
249		}
250		}
251

martin-georgiev / postgresql-for-doctrine

Pull Request — main (#391)

parsePostgresArrayManually() B

Complexity

Size

Duplication

Code Coverage

Importance

How to fix Complexity

Long Method

Duplication Side-by-Side

Filter issues like