fix(flow-php/etl): rename_map() breaks on numeric entry names

.github/workflows/job-arrow-extension.yml

@@ -109,7 +109,7 @@ jobs:
 
       - name: Upload to Codecov
         if: ${{ !cancelled() && matrix.php == '8.3' && matrix.os == 'ubuntu-latest' }}
-        uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7
+        uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7.0.0
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           directory: ./var/phpunit/coverage/clover

.github/workflows/job-extension-tests.yml

@@ -53,7 +53,7 @@ jobs:
 
       - name: Upload to Codecov
         if: ${{ !cancelled() && matrix.php-version == '8.3' }}
-        uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7
+        uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7.0.0
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           directory: ./var/phpunit/coverage/clover

.github/workflows/job-flow-php-extension.yml

@@ -106,7 +106,7 @@ jobs:
 
       - name: Upload to Codecov
         if: ${{ !cancelled() && matrix.php == '8.3' && matrix.os == 'ubuntu-latest' }}
-        uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7
+        uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7.0.0
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           directory: ./var/phpunit/coverage/clover

.github/workflows/job-phpunit-telemetry-tests.yml

@@ -72,7 +72,7 @@ jobs:
 
       - name: Upload to Codecov
         if: ${{ !cancelled() && matrix.php-version == '8.3' && matrix.phpunit-version == '11' }}
-        uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7
+        uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f # v7.0.0
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
           directory: ./var/phpunit/coverage/clover

src/adapter/etl-adapter-csv/src/Flow/ETL/Adapter/CSV/CSVExtractor.php

@@ -32,6 +32,7 @@
 
 use function array_diff;
 use function array_keys;
+use function array_map;
 use function array_values;
 use function iterator_to_array;
 use function sprintf;
@@ -121,7 +122,10 @@ public function extract(FlowContext $context): Generator
 
             foreach ($reader->batches($source, $batchSize) as $rawBatch) {
                 if ($columns === null) {
-                    $columns = array_values(array_diff(array_keys($rawBatch[0]->values), $tail));
+                    $columns = array_map(
+                        strval(...),
+                        array_values(array_diff(array_keys($rawBatch[0]->values), $tail)),
+                    );
 
                     if (
                         $this->schema === null

src/adapter/etl-adapter-parquet/src/Flow/ETL/Adapter/Parquet/ParquetEncoder.php

@@ -14,7 +14,7 @@
 use function array_key_exists;
 
 /**
- * @implements Encoder<array<string, mixed>>
+ * @implements Encoder<array<array-key, mixed>>
  */
 final class ParquetEncoder implements Encoder
 {

src/adapter/etl-adapter-parquet/src/Flow/ETL/Adapter/Parquet/ParquetLoader.php

@@ -218,7 +218,9 @@ private function openWriter(DestinationStream $stream): Writer
 
     private function schema(): Schema
     {
-        return ($this->schema ?? $this->inferredSchema ?? throw new RuntimeException(
+        return ($this->schema
+        ?? $this->inferredSchema
+        ?? throw new RuntimeException(
             'Schema has not been inferred yet. Load at least one batch of rows first.',
         ))->gracefulRemove(...$this->router->droppedNames());
     }

src/adapter/etl-adapter-parquet/tests/Flow/ETL/Adapter/Parquet/Tests/Context/ParquetFilesContext.php

@@ -33,7 +33,7 @@ public static function columnNames(Filesystem $filesystem, string $uri): array
     }
 
     /**
-     * @return list<array<string, mixed>>
+     * @return list<array<array-key, mixed>>
      */
     public static function values(Filesystem $filesystem, string $uri): array
     {

src/core/etl/src/Flow/ETL/DSL/functions.php

@@ -575,7 +575,7 @@ function rename_replace(string|array $search, string|array $replace): RenameRepl
 }
 
 /**
- * @param array<string, string> $renames Map of old_name => new_name
+ * @param array<array-key, string> $renames Map of old_name => new_name
  */
 #[DocumentationDSL(module: Module::CORE, type: DSLType::TRANSFORMER)]
 function rename_map(array $renames): RenameMapEntryStrategy
@@ -584,7 +584,7 @@ function rename_map(array $renames): RenameMapEntryStrategy
 }
 
 /**
- * @param array<string, mixed> $values
+ * @param array<array-key, mixed> $values
  */
 #[DocumentationDSL(module: Module::CORE, type: DSLType::DATA_FRAME)]
 function row(array $values): Row

src/core/etl/src/Flow/ETL/Extractor/FileConstants.php

@@ -19,9 +19,9 @@ public function __construct(
     ) {}
 
     /**
-     * @param array<string, mixed> $row
+     * @param array<array-key, mixed> $row
      *
-     * @return array<string, mixed>
+     * @return array<array-key, mixed>
      */
     public function fill(array $row): array
     {

src/core/etl/src/Flow/ETL/Extractor/PartitionColumns.php

@@ -119,11 +119,11 @@ public function apply(Rows $rows, array $names, PartitionTypes $types = new Part
     }
 
     /**
-     * @param array<string, bool> $names
-     * @param array<string, mixed> $values partition name => value, for the path this row came from
-     * @param array<string, mixed> $row
+     * @param array<array-key, bool> $names
+     * @param array<array-key, mixed> $values partition name => value, for the path this row came from
+     * @param array<array-key, mixed> $row
      *
-     * @return array<string, mixed>
+     * @return array<array-key, mixed>
      */
     public function fill(array $row, array $names, array $values): array
     {

src/core/etl/src/Flow/ETL/Join/HashJoin/RowMerger.php

@@ -15,17 +15,17 @@
 final class RowMerger
 {
     /**
-     * @var array<string, true>
+     * @var array<array-key, true>
      */
     private readonly array $dropLeft;
 
     /**
-     * @var array<string, true>
+     * @var array<array-key, true>
      */
     private readonly array $dropRight;
 
     /**
-     * @var array<string, array{array<string, true>, array<string, true>, array<string, string>}>
+     * @var array<string, array{array<array-key, true>, array<array-key, true>, array<array-key, string>}>
      */
     private array $plans = [];
 
@@ -89,12 +89,12 @@ public function merge(Row $left, Row $right): Row
     }
 
     /**
-     * @param array<string> $leftNames
-     * @param array<string> $rightNames
+     * @param list<array-key> $leftNames
+     * @param list<array-key> $rightNames
      *
      * @throws DuplicatedEntriesException
      *
-     * @return array{array<string, true>, array<string, true>, array<string, string>}
+     * @return array{array<array-key, true>, array<array-key, true>, array<array-key, string>}
      */
     private function plan(array $leftNames, array $rightNames): array
     {
@@ -121,9 +121,11 @@ private function plan(array $leftNames, array $rightNames): array
             }
 
             $keepRight[$name] = true;
-            $outputName = $this->prefix === '' ? $name : $this->prefix . $name;
 
-            if ($this->prefix !== '') {
+            if ($this->prefix === '') {
+                $outputName = $name;
+            } else {
+                $outputName = $this->prefix . $name;
                 $renames[$name] = $outputName;
             }
 

src/core/etl/src/Flow/ETL/Row.php

@@ -23,7 +23,7 @@
 final readonly class Row
 {
     /**
-     * @param array<string, mixed> $values storage keyed by name; the Schema defines column order (Rows::schema()->references())
+     * @param array<array-key, mixed> $values storage keyed by name; the Schema defines column order (Rows::schema()->references())
      */
     public function __construct(
         private array $values,
@@ -129,7 +129,7 @@ public function hash(Schema $schema, Algorithm $algorithm = new NativePHPHash())
      */
     public function names(): array
     {
-        return array_keys($this->values);
+        return array_map(strval(...), array_keys($this->values));
     }
 
     /**
@@ -152,7 +152,7 @@ public function toArray(bool $withKeys = true): array
     }
 
     /**
-     * @return array<string, mixed>
+     * @return array<array-key, mixed>
      */
     public function values(): array
     {
@@ -195,7 +195,7 @@ private function conform(Schema $schema, bool $checkValues): self
         if ($taken !== count($this->values)) {
             foreach ($this->values as $name => $_) {
                 if (!array_key_exists($name, $definitions)) {
-                    throw ColumnMismatchException::unexpectedColumn($name);
+                    throw ColumnMismatchException::unexpectedColumn((string) $name);
                 }
             }
         }

src/core/etl/src/Flow/ETL/Row/Comparator/NativeComparator.php

@@ -9,11 +9,13 @@
 use Flow\ETL\Row\TypedValueComparator;
 use Flow\ETL\Schema;
 
+use function array_keys;
+
 final class NativeComparator implements Comparator
 {
     public function equals(Row $row, Row $nextRow, Schema $schema): bool
     {
-        if ($row->names() !== $nextRow->names()) {
+        if (array_keys($row->values()) !== array_keys($nextRow->values())) {
             return false;
         }
 

src/core/etl/src/Flow/ETL/Row/InferredBatch.php

@@ -36,7 +36,6 @@ public function of(array $batch): Rows
                 // RawRowValues::$values is declared array<string, mixed>, but PHP hands a numeric
                 // column name back as an INT key - a pivot names its columns by their own values.
                 // definition_from_type() takes Reference|string, so the cast is not redundant at runtime.
-                // @mago-ignore analysis:redundant-cast
                 $name = (string) $name;
                 $definition = definition_from_type($name, $this->typeDetector->detectType($value), $value === null);
 
@@ -60,7 +59,6 @@ public function of(array $batch): Rows
             /** @var mixed $value */
             foreach ($rowValues->values as $name => $value) {
                 // see above - an int key here would miss the column the schema declares as "5"
-                // @mago-ignore analysis:redundant-cast
                 $name = (string) $name;
                 $definition = $schema->get($name);
 

src/core/etl/src/Flow/ETL/Row/RawRowValues.php

@@ -9,7 +9,7 @@
 final readonly class RawRowValues
 {
     /**
-     * @param array<string, mixed> $values
+     * @param array<array-key, mixed> $values
      * @param array<array-key, Metadata> $metadata
      */
     public function __construct(

src/core/etl/src/Flow/ETL/Row/RowRenaming.php

@@ -9,14 +9,14 @@
 final readonly class RowRenaming
 {
     /**
-     * @param array<string, string> $renames current_name => new_name
+     * @param array<array-key, string> $renames current_name => new_name
      */
     private function __construct(
         private array $renames,
     ) {}
 
     /**
-     * @param array<string, string> $renames current_name => new_name
+     * @param array<array-key, string> $renames current_name => new_name
      */
     public static function of(array $renames): self
     {

src/core/etl/src/Flow/ETL/Rows.php

@@ -118,7 +118,7 @@ public static function conformed(Schema $schema, array $rows): self
     }
 
     /**
-     * @return array{schema: Schema, rows: list<array<string, mixed>>}
+     * @return array{schema: Schema, rows: list<array<array-key, mixed>>}
      */
     public function __serialize(): array
     {
@@ -149,7 +149,7 @@ public function __serialize(): array
     }
 
     /**
-     * @param array{schema: Schema, rows: list<array<string, mixed>>} $data
+     * @param array{schema: Schema, rows: list<array<array-key, mixed>>} $data
      */
     public function __unserialize(array $data): void
     {

src/core/etl/src/Flow/ETL/Transformer/Rename/RenameEntryStrategy.php

@@ -9,7 +9,7 @@
 interface RenameEntryStrategy
 {
     /**
-     * @return array<string, string> map of current_name => new_name, empty when nothing is renamed
+     * @return array<array-key, string> map of current_name => new_name; PHP casts a numeric name to an int key, empty when nothing is renamed
      */
     public function renames(Schema $schema): array;
 }

src/core/etl/src/Flow/ETL/Transformer/Rename/RenameMapEntryStrategy.php

@@ -12,7 +12,7 @@
 final readonly class RenameMapEntryStrategy implements RenameEntryStrategy
 {
     /**
-     * @param array<string, string> $renames Map of old_name => new_name
+     * @param array<array-key, string> $renames Map of old_name => new_name
      */
     public function __construct(
         private array $renames,

src/core/etl/src/Flow/ETL/Transformer/RenameEachEntryTransformer.php

@@ -16,7 +16,6 @@
 use Throwable;
 
 use function array_search;
-use function is_string;
 
 final readonly class RenameEachEntryTransformer implements Transformer
 {
@@ -40,7 +39,7 @@ public function bind(Schema $input): BoundStep
 
         foreach ($this->strategies as $strategy) {
             foreach ($strategy->renames($output) as $from => $to) {
-                $output = $output->rename($from, $to);
+                $output = $output->rename((string) $from, $to);
             }
         }
 
@@ -57,13 +56,16 @@ public function transform(Rows $rows, FlowContext $context): Rows
 
             foreach ($this->strategies as $strategy) {
                 foreach ($strategy->renames($schema) as $from => $to) {
+                    // a numeric entry name arrives as int - PHP casts numeric-string array keys
+                    $from = (string) $from;
+
                     $schema = $schema->rename($from, $to);
 
                     // strategies chain, so a later one renames what an earlier one produced; the map
                     // has to stay keyed by the row's original name or the projection lands short
                     $original = array_search($from, $renames, true);
 
-                    $renames[is_string($original) ? $original : $from] = $to;
+                    $renames[false === $original ? $from : $original] = $to;
                 }
             }
 

src/core/etl/src/Flow/ETL/Transformer/UnpackedColumns.php

@@ -28,10 +28,10 @@ public function of(Schema $base, string $prefix, Schema $declared): Schema
     }
 
     /**
-     * @param array<string, mixed> $values
+     * @param array<array-key, mixed> $values
      * @param array<array-key, mixed> $payload
      *
-     * @return array<string, mixed>
+     * @return array<array-key, mixed>
      */
     public function values(array $values, string $prefix, Schema $declared, array $payload): array
     {

src/core/etl/src/Flow/Floe/FloeLoader.php

@@ -159,9 +159,9 @@ private function closeWriters(): void
     private function openWriter(DestinationStream $stream, FlowContext $context): FloeWriter
     {
         $declared =
-            $this->schema ?? $this->inferredSchema ?? throw new FloeException(
-                'Floe loader has no schema to write with',
-            );
+            $this->schema
+            ?? $this->inferredSchema
+            ?? throw new FloeException('Floe loader has no schema to write with');
 
         // partitionBy() keeps the partition columns in the path rather than the body, so the file
         // schema must not declare them either. Every other declared column stays - a batch that does

src/core/etl/tests/Flow/ETL/Tests/Integration/DataFrame/DisplayTest.php

@@ -91,7 +91,6 @@ enum_schema('enum', BackedStringEnum::class),
                             xml_schema('xml'),
                         ),
                         // PHP casts the numeric column name "100" to an int array key
-                        // @mago-ignore analysis:possibly-invalid-argument
                         row([
                             'id' => 1234,
                             'price' => 123.45,
@@ -159,7 +158,6 @@ public function extract(FlowContext $context): Generator
                                 string_schema('group'),
                             ),
                             // PHP casts the numeric column name "100" to an int array key
-                            // @mago-ignore analysis:possibly-invalid-argument
                             row([
                                 'id' => 1234,
                                 'price' => 123.45,
@@ -182,7 +180,6 @@ public function extract(FlowContext $context): Generator
                                 string_schema('group'),
                             ),
                             // PHP casts the numeric column name "100" to an int array key
-                            // @mago-ignore analysis:possibly-invalid-argument
                             row([
                                 'id' => 1234,
                                 'price' => 123.45,