From 2761773b3dddf62b9ccfb3919b7b7c935935408a Mon Sep 17 00:00:00 2001
From: Bill Blume <bill.blume@mobileforcesoftware.com>
Date: Mon, 31 Jul 2017 11:36:54 -0700
Subject: [PATCH] Merge data-validation collections to reduce the final file
 size

Closes #131
Closes #193
---
 CHANGELOG.md                                 |  2 +
 src/PhpSpreadsheet/Cell.php                  | 80 ++++++++++++++++++++
 src/PhpSpreadsheet/Writer/Xlsx/Worksheet.php |  1 +
 tests/PhpSpreadsheetTests/CellTest.php       | 16 ++++
 tests/data/CellMergeRangesInCollection.php   | 58 ++++++++++++++
 5 files changed, 157 insertions(+)
 create mode 100644 tests/data/CellMergeRangesInCollection.php

diff --git a/CHANGELOG.md b/CHANGELOG.md
index f7179613..94204fa6 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -10,6 +10,8 @@ and this project adheres to [Semantic Versioning](http://semver.org/).
 
 ### Changed
 
+- Merge data-validations to reduce written worksheet size - @billblume [#131](https://github.com/PHPOffice/PhpSpreadSheet/issues/131)
+
 ### Fixed
 
 
diff --git a/src/PhpSpreadsheet/Cell.php b/src/PhpSpreadsheet/Cell.php
index 0263853d..2ed9b47f 100644
--- a/src/PhpSpreadsheet/Cell.php
+++ b/src/PhpSpreadsheet/Cell.php
@@ -948,6 +948,86 @@ class Cell
         return array_values($sortKeys);
     }
 
+    /**
+     * Convert an associative array of single cell coordinates to values to an associative array
+     * of cell ranges to values.  Only adjacent cell coordinates with the same
+     * value will be merged.  If the value is an object, it must implement the method getHashCode().
+     *
+     * For example, this function converts:
+     *
+     *    [ 'A1' => 'x', 'A2' => 'x', 'A3' => 'x', 'A4' => 'y' ]
+     *
+     * to:
+     *
+     *    [ 'A1:A3' => 'x', 'A4' => 'y' ]
+     *
+     * @param array $pCoordCollection associative array mapping coordinates to values
+     *
+     * @return array associative array mapping coordinate ranges to valuea
+     */
+    public static function mergeRangesInCollection(array $pCoordCollection)
+    {
+        $hashedValues = [];
+
+        foreach ($pCoordCollection as $coord => $value) {
+            list($column, $row) = self::coordinateFromString($coord);
+            $row = (int) (ltrim($row, '$'));
+            $hashCode = $column . '-' . (is_object($value) ? $value->getHashCode() : $value);
+
+            if (!isset($hashedValues[$hashCode])) {
+                $hashedValues[$hashCode] = (object) [
+                    'value' => $value,
+                    'col' => $column,
+                    'rows' => [$row],
+                ];
+            } else {
+                $hashedValues[$hashCode]->rows[] = $row;
+            }
+        }
+
+        $mergedCoordCollection = [];
+        ksort($hashedValues);
+
+        foreach ($hashedValues as $hashedValue) {
+            sort($hashedValue->rows);
+            $rowStart = null;
+            $rowEnd = null;
+            $ranges = [];
+
+            foreach ($hashedValue->rows as $row) {
+                if ($rowStart === null) {
+                    $rowStart = $row;
+                    $rowEnd = $row;
+                } elseif ($rowEnd === $row - 1) {
+                    $rowEnd = $row;
+                } else {
+                    if ($rowStart == $rowEnd) {
+                        $ranges[] = $hashedValue->col . $rowStart;
+                    } else {
+                        $ranges[] = $hashedValue->col . $rowStart . ':' . $hashedValue->col . $rowEnd;
+                    }
+
+                    $rowStart = $row;
+                    $rowEnd = $row;
+                }
+            }
+
+            if ($rowStart !== null) {
+                if ($rowStart == $rowEnd) {
+                    $ranges[] = $hashedValue->col . $rowStart;
+                } else {
+                    $ranges[] = $hashedValue->col . $rowStart . ':' . $hashedValue->col . $rowEnd;
+                }
+            }
+
+            foreach ($ranges as $range) {
+                $mergedCoordCollection[$range] = $hashedValue->value;
+            }
+        }
+
+        return $mergedCoordCollection;
+    }
+
     /**
      * Compare 2 cells.
      *
diff --git a/src/PhpSpreadsheet/Writer/Xlsx/Worksheet.php b/src/PhpSpreadsheet/Writer/Xlsx/Worksheet.php
index a1a97865..dfaea5a0 100644
--- a/src/PhpSpreadsheet/Writer/Xlsx/Worksheet.php
+++ b/src/PhpSpreadsheet/Writer/Xlsx/Worksheet.php
@@ -565,6 +565,7 @@ class Worksheet extends WriterPart
 
         // Write data validations?
         if (!empty($dataValidationCollection)) {
+            $dataValidationCollection = Cell::mergeRangesInCollection($dataValidationCollection);
             $objWriter->startElement('dataValidations');
             $objWriter->writeAttribute('count', count($dataValidationCollection));
 
diff --git a/tests/PhpSpreadsheetTests/CellTest.php b/tests/PhpSpreadsheetTests/CellTest.php
index 118978f9..3a7d429e 100644
--- a/tests/PhpSpreadsheetTests/CellTest.php
+++ b/tests/PhpSpreadsheetTests/CellTest.php
@@ -300,4 +300,20 @@ class CellTest extends PHPUnit_Framework_TestCase
     {
         return require 'data/CellExtractAllCellReferencesInRange.php';
     }
+
+    /**
+     * @dataProvider providerMergeRangesInCollection
+     *
+     * @param mixed $expectedResult
+     */
+    public function testMergeRangesInCollection($expectedResult, ...$args)
+    {
+        $result = Cell::mergeRangesInCollection(...$args);
+        $this->assertEquals($expectedResult, $result);
+    }
+
+    public function providerMergeRangesInCollection()
+    {
+        return require 'data/CellMergeRangesInCollection.php';
+    }
 }
diff --git a/tests/data/CellMergeRangesInCollection.php b/tests/data/CellMergeRangesInCollection.php
new file mode 100644
index 00000000..9cb71125
--- /dev/null
+++ b/tests/data/CellMergeRangesInCollection.php
@@ -0,0 +1,58 @@
+<?php
+
+return [
+    [
+        [
+            'A1:A3' => 'x',
+            'A4' => 'y',
+        ],
+        [
+            'A1' => 'x',
+            'A2' => 'x',
+            'A3' => 'x',
+            'A4' => 'y',
+        ],
+    ],
+    [
+        [
+            'A1:A4' => 'x',
+            'A6:A7' => 'x',
+            'A9' => 'x',
+        ],
+        [
+            'A7' => 'x',
+            'A1' => 'x',
+            'A4' => 'x',
+            'A6' => 'x',
+            'A2' => 'x',
+            'A9' => 'x',
+            'A3' => 'x',
+        ],
+    ],
+    [
+        [
+            'A1:A3' => 'x',
+            'B1:B3' => 'x',
+        ],
+        [
+            'A1' => 'x',
+            'B3' => 'x',
+            'A2' => 'x',
+            'B2' => 'x',
+            'A3' => 'x',
+            'B1' => 'x',
+        ],
+    ],
+    [
+        [
+            'A1' => 'x',
+            'A2' => 'y',
+            'A3' => 'z',
+        ],
+        [
+            'A1' => 'x',
+            'A2' => 'y',
+            'A3' => 'z',
+        ],
+    ],
+];