CSV Writer Allow Varying Number of Columns

Supersedes PR #1415 by @AndrewMonty, which went stale in May 2020, and which is not directly usable due to changes between now and then. Fix #1414, which also went stale; I will remove the stale status and reopen the issue pending the merging of this PR.

Add an option to CSV Writer so that it writes the cells for a row only through the highest data column used in the row, rather than through the highest data column used in the worksheet.

Author: oleibman

CHANGELOG.md

@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org).
 ### Added
 
 - Xlsx Reader Optionally Ignore Rows With No Cells. [Issue #3982](https://github.com/PHPOffice/PhpSpreadsheet/issues/3982) [PR #4035](https://github.com/PHPOffice/PhpSpreadsheet/pull/4035)
+- Option for CSV output file to have varying numbers of columns for each row. [Issue #1415](https://github.com/PHPOffice/PhpSpreadsheet/issues/1415) [PR #4076](https://github.com/PHPOffice/PhpSpreadsheet/pull/4076)
 
 ### Changed
 

docs/topics/reading-and-writing-to-file.md

@@ -679,6 +679,18 @@ $writer->setOutputEncoding('SJIS-WIN');
 $writer->save("05featuredemo.csv");
 ```
 
+#### Writing CSV files with varying numbers of columns
+
+A CSV file can have a different number of columns in each row. This
+differs from the default behavior when saving as a .csv in Excel, but
+can be enabled in PhpSpreadsheet by using the following code:
+
+``` php
+$writer = new \PhpOffice\PhpSpreadsheet\Writer\Csv($spreadsheet);
+$writer->setVariableColumns(true);
+$writer->save("05featuredemo.csv");
+```
+
 #### Decimal and thousands separators
 
 If the worksheet you are exporting contains numbers with decimal or

src/PhpSpreadsheet/Writer/Csv.php

@@ -3,6 +3,7 @@
 namespace PhpOffice\PhpSpreadsheet\Writer;
 
 use PhpOffice\PhpSpreadsheet\Calculation\Calculation;
+use PhpOffice\PhpSpreadsheet\Cell\Coordinate;
 use PhpOffice\PhpSpreadsheet\Spreadsheet;
 use Stringable;
 
@@ -54,6 +55,13 @@ class Csv extends BaseWriter
      */
     private string $outputEncoding = '';
 
+    /**
+     * Whether number of columns should be allowed to vary
+     * between rows, or use a fixed range based on the max
+     * column overall.
+     */
+    private bool $variableColumns = false;
+
     /**
      * Create a new CSV.
      */
@@ -105,7 +113,17 @@ public function save($filename, int $flags = 0): void
         $maxRow = $sheet->getHighestDataRow();
 
         // Write rows to file
+        $row = 0;
         foreach ($sheet->rangeToArrayYieldRows("A1:$maxCol$maxRow", '', $this->preCalculateFormulas) as $cellsArray) {
+            ++$row;
+            if ($this->variableColumns) {
+                $column = $sheet->getHighestDataColumn($row);
+                if ($column === 'A' && !$sheet->cellExists("A$row")) {
+                    $cellsArray = [];
+                } else {
+                    array_splice($cellsArray, Coordinate::columnIndexFromString($column));
+                }
+            }
             $this->writeLine($this->fileHandle, $cellsArray);
         }
 
@@ -303,4 +321,26 @@ private function writeLine($fileHandle, array $values): void
         }
         fwrite($fileHandle, $line);
     }
+
+    /**
+     * Get whether number of columns should be allowed to vary
+     * between rows, or use a fixed range based on the max
+     * column overall.
+     */
+    public function getVariableColumns(): bool
+    {
+        return $this->variableColumns;
+    }
+
+    /**
+     * Set whether number of columns should be allowed to vary
+     * between rows, or use a fixed range based on the max
+     * column overall.
+     */
+    public function setVariableColumns(bool $pValue): self
+    {
+        $this->variableColumns = $pValue;
+
+        return $this;
+    }
 }

tests/PhpSpreadsheetTests/Writer/Csv/VariableColumnsTest.php

@@ -0,0 +1,81 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PhpOffice\PhpSpreadsheetTests\Writer\Csv;
+
+use PhpOffice\PhpSpreadsheet\Shared\File;
+use PhpOffice\PhpSpreadsheet\Spreadsheet;
+use PhpOffice\PhpSpreadsheet\Writer\Csv;
+use PHPUnit\Framework\TestCase;
+
+class VariableColumnsTest extends TestCase
+{
+    public function testVariableColumns(): void
+    {
+        $spreadsheet = new Spreadsheet();
+        $sheet = $spreadsheet->getActiveSheet();
+        $sheet->fromArray(
+            [
+                [1, 2, 3, 4],
+                [1, 2],
+                [1, 2, 3, 4, 5],
+                [],
+                [1],
+                [1, 2, 3],
+            ]
+        );
+
+        $filename = File::temporaryFilename();
+        $writer = new Csv($spreadsheet);
+        $writer->setVariableColumns(true);
+        $writer->save($filename);
+
+        $contents = (string) file_get_contents($filename);
+        unlink($filename);
+        $spreadsheet->disconnectWorksheets();
+
+        $rows = explode(PHP_EOL, $contents);
+
+        self::assertSame('"1","2","3","4"', $rows[0]);
+        self::assertSame('"1","2"', $rows[1]);
+        self::assertSame('"1","2","3","4","5"', $rows[2]);
+        self::assertSame('', $rows[3]);
+        self::assertSame('"1"', $rows[4]);
+        self::assertSame('"1","2","3"', $rows[5]);
+    }
+
+    public function testFixedColumns(): void
+    {
+        $spreadsheet = new Spreadsheet();
+        $sheet = $spreadsheet->getActiveSheet();
+        $sheet->fromArray(
+            [
+                [1, 2, 3, 4],
+                [1, 2],
+                [1, 2, 3, 4, 5],
+                [],
+                [1],
+                [1, 2, 3],
+            ]
+        );
+
+        $filename = File::temporaryFilename();
+        $writer = new Csv($spreadsheet);
+        self::assertFalse($writer->getVariableColumns());
+        $writer->save($filename);
+
+        $contents = (string) file_get_contents($filename);
+        unlink($filename);
+        $spreadsheet->disconnectWorksheets();
+
+        $rows = explode(PHP_EOL, $contents);
+
+        self::assertSame('"1","2","3","4",""', $rows[0]);
+        self::assertSame('"1","2","","",""', $rows[1]);
+        self::assertSame('"1","2","3","4","5"', $rows[2]);
+        self::assertSame('"","","","",""', $rows[3]);
+        self::assertSame('"1","","","",""', $rows[4]);
+        self::assertSame('"1","2","3","",""', $rows[5]);
+    }
+}