Merge pull request #241 from toddlucas/master

Fix XSSFExcelExtractor to emit empty cells
nissl-lab · Jan 11, 2020 · e918a6e · e918a6e
2 parents 1304e24 + 6b13de1
commit e918a6e
Show file tree

Hide file tree

Showing 2 changed files with 35 additions and 2 deletions.
diff --git a/ooxml/XSSF/Extractor/XSSFExcelExtractor.cs b/ooxml/XSSF/Extractor/XSSFExcelExtractor.cs
@@ -192,7 +192,7 @@ public override String Text
                         IRow row = (IRow)rawR;
                         IEnumerator ri =row.GetEnumerator();
                         bool firsttime=true;
-                        while (ri.MoveNext())
+                        for (int j = 0; j < row.LastCellNum; j++)
                         {
                             if (!firsttime)
                             {
@@ -202,7 +202,9 @@ public override String Text
                             {
                                 firsttime = false;
                             }
-                            ICell cell = (ICell)ri.Current;
+                            ICell cell = (ICell)row.GetCell(j);
+                            if (cell == null)
+                                continue;
 
                             // Is it a formula one?
                             if (cell.CellType == CellType.Formula)

diff --git a/testcases/ooxml/XSSF/Extractor/TestXSSFExcelExtractor.cs b/testcases/ooxml/XSSF/Extractor/TestXSSFExcelExtractor.cs
@@ -212,6 +212,37 @@ public void TestInlineStrings()
             Assert.IsTrue(text.Contains("A5-A$2"), "Unable to find expected word in text\n" + text);
             extractor.Close();
         }
+
+        [Test]
+        public void TestEmptyCells()
+        {
+            XSSFExcelExtractor extractor = GetExtractor("SimpleNormal.xlsx");
+
+            String text = extractor.Text;
+            Assert.IsTrue(text.Length > 0);
+
+            // This sheet demonstrates the preservation of empty cells, as
+            // signified by sequential \t characters.
+            Assert.AreEqual(
+                // Sheet 1
+                "Sheet1\n" + 
+                "test\t\t1\n" + 
+                "test 2\t\t2\n" + 
+                "\t\t3\n" + 
+                "\t\t4\n" + 
+                "\t\t5\n" + 
+                "\t\t6\n" + 
+                // Sheet 2
+                "Sheet Number 2\n" + 
+                "This is sheet 2\n" + 
+                "Stuff\n" + 
+                "1\t2\t3\t4\t5\t6\n" + 
+                "1/1/90\n" + 
+                "10\t\t3\n", 
+                text);
+
+            extractor.Close();
+        }
     }
 }