make int parsing more tolerant of stray whitespace

pjfanning · Sep 12, 2023 · 0802856 · 0802856
1 parent 1270672
commit 0802856
Show file tree

Hide file tree

Showing 7 changed files with 46 additions and 13 deletions.
diff --git a/src/main/java/com/github/pjfanning/xlsx/impl/BaseEvaluationWorkbook.java b/src/main/java/com/github/pjfanning/xlsx/impl/BaseEvaluationWorkbook.java
@@ -15,6 +15,8 @@
 import org.apache.poi.xssf.usermodel.XSSFName;
 import org.apache.poi.xssf.usermodel.XSSFTable;
 
+import static com.github.pjfanning.xlsx.impl.NumberUtil.parseInt;
+
 /**
  * Copied from POI BaseXSSFEvaluationWorkbook but a lot of stuff is removed because it is not easy
  * or impossible to support in excel-streaming-reader
@@ -78,7 +80,7 @@ private int resolveBookIndex(String bookName) {
 
     // Is it already in numeric form?
     try {
-      return Integer.parseInt(bookName);
+      return parseInt(bookName);
     } catch (NumberFormatException e) {}
 
     // Not properly referenced

diff --git a/src/main/java/com/github/pjfanning/xlsx/impl/NumberUtil.java b/src/main/java/com/github/pjfanning/xlsx/impl/NumberUtil.java
@@ -0,0 +1,12 @@
+package com.github.pjfanning.xlsx.impl;
+
+final class NumberUtil {
+
+  static int parseInt(final String s) {
+    return Integer.parseInt(s.trim());
+  }
+
+  static double parseDouble(final String s) {
+    return Double.parseDouble(s.trim());
+  }
+}
diff --git a/src/main/java/com/github/pjfanning/xlsx/impl/StreamingCell.java b/src/main/java/com/github/pjfanning/xlsx/impl/StreamingCell.java
@@ -13,6 +13,8 @@
 import java.util.Calendar;
 import java.util.Date;
 
+import static com.github.pjfanning.xlsx.impl.NumberUtil.parseDouble;
+
 public class StreamingCell implements Cell {
 
   private static final Supplier NULL_SUPPLIER = () -> null;
@@ -197,7 +199,7 @@ public double getNumericCellValue() {
         }
       }
     }
-    return rawContents == null ? 0.0 : Double.parseDouble(rawContents);
+    return rawContents == null ? 0.0 : parseDouble(rawContents);
   }
 
   /**

diff --git a/src/main/java/com/github/pjfanning/xlsx/impl/StreamingRowIterator.java b/src/main/java/com/github/pjfanning/xlsx/impl/StreamingRowIterator.java
@@ -38,6 +38,9 @@
 import java.time.LocalDateTime;
 import java.util.*;
 
+import static com.github.pjfanning.xlsx.impl.NumberUtil.parseDouble;
+import static com.github.pjfanning.xlsx.impl.NumberUtil.parseInt;
+
 class StreamingRowIterator implements CloseableIterator<Row> {
   private static final Logger LOG = LoggerFactory.getLogger(StreamingRowIterator.class);
   private static final QName QNAME_HIDDEN = QName.valueOf("hidden");
@@ -144,7 +147,7 @@ && isSpreadsheetTag(event.asStartElement().getName())) {
         Attribute rowNumAttr = startElement.getAttributeByName(QNAME_R);
         int rowIndex = currentRowNum;
         if (rowNumAttr != null) {
-          rowIndex = Integer.parseInt(rowNumAttr.getValue()) - 1;
+          rowIndex = parseInt(rowNumAttr.getValue()) - 1;
           currentRowNum = rowIndex;
         }
         Attribute isHiddenAttr = startElement.getAttributeByName(QNAME_HIDDEN);
@@ -166,7 +169,7 @@ && isSpreadsheetTag(event.asStartElement().getName())) {
           if (styleAttr != null) {
             String indexStr = styleAttr.getValue();
             try {
-              int index = Integer.parseInt(indexStr);
+              final int index = parseInt(indexStr);
               currentRow.setRowStyle(stylesTable.getStyleAt(index));
             } catch (NumberFormatException nfe) {
               LOG.warn("Ignoring invalid row style index {}", indexStr);
@@ -189,8 +192,8 @@ && isSpreadsheetTag(event.asStartElement().getName())) {
         if (isHidden || width >= 0) {
           Attribute minAttr = startElement.getAttributeByName(QNAME_MIN);
           Attribute maxAttr = startElement.getAttributeByName(QNAME_MAX);
-          int min = Integer.parseInt(minAttr.getValue()) - 1;
-          int max = Integer.parseInt(maxAttr.getValue()) - 1;
+          int min = parseInt(minAttr.getValue()) - 1;
+          int max = parseInt(maxAttr.getValue()) - 1;
           for (int columnIndex = min; columnIndex <= max; columnIndex++) {
             if (isHidden) hiddenColumns.add(columnIndex);
             if (width >= 0) columnWidths.put(columnIndex, width);
@@ -226,7 +229,7 @@ && isSpreadsheetTag(event.asStartElement().getName())) {
           if (style != null) {
             String indexStr = style.getValue();
             try {
-              int index = Integer.parseInt(indexStr);
+              final int index = Integer.parseInt(indexStr);
               currentCell.setCellStyle(stylesTable.getStyleAt(index));
             } catch (NumberFormatException nfe) {
               LOG.warn("Ignoring invalid style index {}", indexStr);
@@ -249,7 +252,7 @@ && isSpreadsheetTag(event.asStartElement().getName())) {
           for (int i = ref.length() - 1; i >= 0; i--) {
             if (!Character.isDigit(ref.charAt(i))) {
               try {
-                streamingSheetReader.setLastRowNum(Integer.parseInt(ref.substring(i + 1)) - 1);
+                streamingSheetReader.setLastRowNum(parseInt(ref.substring(i + 1)) - 1);
               } catch (NumberFormatException ignore) {
               }
               break;
@@ -467,7 +470,7 @@ private void parseSheetFormatPr(final StartElement startElement) {
     final Attribute baseColWidthAtt = startElement.getAttributeByName(QName.valueOf("baseColWidth"));
     if (baseColWidthAtt != null) {
       try {
-        streamingSheetReader.setBaseColWidth(Integer.parseInt(baseColWidthAtt.getValue()));
+        streamingSheetReader.setBaseColWidth(parseInt(baseColWidthAtt.getValue()));
       } catch (Exception e) {
         LOG.warn("unable to parse baseColWidth {}", baseColWidthAtt.getValue());
       }
@@ -504,7 +507,7 @@ private Supplier getFormatterForType(String type) {
     switch(type) {
       case "s":           //string stored in shared table
         if (!lastContents.isEmpty()) {
-          int idx = Integer.parseInt(lastContents);
+          final int idx = parseInt(lastContents);
           if (!getBuilder().fullFormatRichText() && sst instanceof SharedStringsTableBase) {
             return new LazySupplier<>(() -> ((SharedStringsTableBase)sst).getString(idx));
           }
@@ -525,7 +528,7 @@ private Supplier getFormatterForType(String type) {
           final String currentNumericFormat = currentCell.getNumericFormat();
 
           return new LazySupplier<>(() -> dataFormatter.formatRawCellContents(
-                  Double.parseDouble(lastContents),
+                  parseDouble(lastContents),
                   currentNumericFormatIndex,
                   currentNumericFormat));
 
@@ -580,7 +583,7 @@ private String unformattedContents(Supplier formattedContentSupplier) {
           return formattedContent.toString();
         }
         if (!lastContents.isEmpty()) {
-          int idx = Integer.parseInt(lastContents);
+          final int idx = parseInt(lastContents);
           if (sst == null) throw new NullPointerException("sst is null");
           if (sst instanceof SharedStringsTableBase) {
             return ((SharedStringsTableBase)sst).getString(idx);

diff --git a/src/main/java/com/github/pjfanning/xlsx/impl/StreamingSheetReader.java b/src/main/java/com/github/pjfanning/xlsx/impl/StreamingSheetReader.java
@@ -37,6 +37,8 @@
 import java.io.IOException;
 import java.util.*;
 
+import static com.github.pjfanning.xlsx.impl.NumberUtil.parseInt;
+
 public class StreamingSheetReader implements Iterable<Row> {
   private static final Logger LOG = LoggerFactory.getLogger(StreamingSheetReader.class);
 
@@ -200,7 +202,7 @@ void setFormatString(StartElement startElement, StreamingCell cell) {
 
     if (stylesTable != null) {
       if(cellStyleString != null) {
-        style = stylesTable.getStyleAt(Integer.parseInt(cellStyleString));
+        style = stylesTable.getStyleAt(parseInt(cellStyleString));
       } else if(stylesTable.getNumCellStyles() > 0) {
         style = stylesTable.getStyleAt(0);
       }

diff --git a/src/test/java/com/github/pjfanning/xlsx/StreamingWorkbookTest.java b/src/test/java/com/github/pjfanning/xlsx/StreamingWorkbookTest.java
@@ -866,6 +866,18 @@ public void testSheetReadWrongOrder() throws Exception {
     }
   }
 
+  @Test
+  public void testExtraWhitespace() throws Exception {
+    try (Workbook workbook = openWorkbook("extra_whitespace.xlsx")) {
+      Sheet sheet = workbook.getSheetAt(0);
+      for(Row row : sheet) {
+        for(Cell cell : row) {
+
+        }
+      }
+    }
+  }
+
   private void validateFormatsSheet(Sheet sheet) throws IOException {
     Iterator<Row> rowIterator = sheet.rowIterator();
 

diff --git a/src/test/resources/extra_whitespace.xlsx b/src/test/resources/extra_whitespace.xlsx