public static void generateTrainingDataFromFile(
String
fileLocation) // Requires that the original file had the metadata and requires that this
// file is formated the same in first sheet
{
testDataLL = (LinkedList<String[]>) dataLL.clone();
actualClassifications = (LinkedList<String>) classificationsLL.clone();
FileInputStream file;
try {
file = new FileInputStream(new File(fileLocation));
Workbook excelFile = new HSSFWorkbook(file);
Sheet sheet1 = excelFile.getSheetAt(0); // Data sheet
for (Row row : sheet1) {
String data[] = new String[row.getPhysicalNumberOfCells() - 1];
String classification = "";
int offset =
0; // Used so that we can declare an array of the size of the attributes without the
// classification
for (Cell cell : row) {
int index = cell.getColumnIndex();
if (classificationLocation != index) {
data[index - offset] = cell.toString();
} else {
classification = cell.toString();
offset++;
}
}
// Even though data and classifications are not really used add it onto the end so it is
// still complete for in the event they end up being used in a later version
dataLL.add(data);
classificationsLL.add(classification);
trainingDataLL.add(data);
knownClassifications.add(classification);
// Check to see if we have seen that classification yet
int occurrences = 0;
for (int i = 0; i < classificationTypes.size() && occurrences == 0; i++) {
if (classificationTypes.get(i).compareTo(classification) == 0) {
occurrences = 1;
}
}
if (occurrences == 0) {
classificationTypes.add(classification);
}
}
excelFile.close();
} catch (FileNotFoundException e) {
System.out.println("Error file not found");
System.exit(0);
} catch (IOException e) {
System.out.println("Unable to read file, disk drive may be failing");
e.printStackTrace();
System.exit(0);
}
}
public static List<List<List<String>>> readExcel(File file, Rule rule) {
int start = rule.getStart();
int end = rule.getEnd();
List<List<List<String>>> result = Lists.newArrayList();
Workbook wb;
try {
wb = WorkbookFactory.create(file);
} catch (Exception e) {
throw new ExcelException(e);
}
for (int i = 0; i < wb.getNumberOfSheets(); i++) {
Sheet sheet = wb.getSheetAt(i);
List<List<String>> sheetList = Lists.newArrayList();
int rows = sheet.getLastRowNum();
if (start <= sheet.getFirstRowNum()) {
start = sheet.getFirstRowNum();
}
if (end >= rows) {
end = rows;
} else if (end <= 0) {
end = rows + end;
}
for (int rowIndex = start; rowIndex <= end; rowIndex++) {
Row row = sheet.getRow(rowIndex);
List<String> columns = Lists.newArrayList();
int cellNum = row.getLastCellNum();
System.out.println(row.getLastCellNum());
System.out.println(row.getPhysicalNumberOfCells());
for (int cellIndex = row.getFirstCellNum(); cellIndex < cellNum; cellIndex++) {
Cell cell = row.getCell(cellIndex);
int cellType = cell.getCellType();
String column = "";
switch (cellType) {
case Cell.CELL_TYPE_NUMERIC:
// DecimalFormat format = new DecimalFormat();
// format.setGroupingUsed(false);
column = String.valueOf(cell.getDateCellValue());
break;
case Cell.CELL_TYPE_STRING:
column = cell.getStringCellValue();
break;
case Cell.CELL_TYPE_BOOLEAN:
column = cell.getBooleanCellValue() + "";
break;
case Cell.CELL_TYPE_FORMULA:
column = cell.getCellFormula();
break;
case Cell.CELL_TYPE_ERROR:
case Cell.CELL_TYPE_BLANK:
column = " ";
break;
default:
}
columns.add(column.trim());
}
List<Boolean> rowFilterFlagList = Lists.newArrayList();
List<RowFilter> rowFilterList = Lists.newArrayList();
for (int k = 0; k < rowFilterList.size(); k++) {
RowFilter rowFilter = rowFilterList.get(k);
rowFilterFlagList.add(rowFilter.doFilter(rowIndex, columns));
}
if (!rowFilterFlagList.contains(false)) {
sheetList.add(columns);
}
}
result.add(sheetList);
}
return result;
}
/**
* 将excel内的内容读取到xml文件中，并添加dtd验证
*
* @param xmlFile
* @param sheetNum
* @return 1代表成功，0失败，-1超过最大sheet,2跳过当前失败的xml
*/
public int excelToXml(String xmlFile, int sheetNum) {
if (sheetNum >= workBook.getNumberOfSheets()) return -1;
else sheet = workBook.getSheetAt(sheetNum);
xmlFile = xmlFile + ".xml";
try {
Document document = DocumentHelper.createDocument();
// 使用sheet名称命名跟节点
String rootName = sheet.getSheetName().replaceAll(" ", "");
Element root = document.addElement(rootName);
// 添加dtd文件说明
DocumentType documentType = new DOMDocumentType();
documentType.setElementName(rootName);
List<ElementDecl> declList = new ArrayList<>();
declList.add(new ElementDecl(rootName, "(row*)"));
// 判断sheet是否为空,为空则不执行任何操作
if (sheet.getRow(0) == null) return 1;
// 遍历sheet第一行,获取元素名称
row = sheet.getRow(0);
String rowString = null;
List<String> pcdataList = new ArrayList<>();
for (int y = 0; y < row.getPhysicalNumberOfCells(); y++) {
Object object = this.getCellValueObject(0, y);
// 判断是否有合并单元格，有的话跳过
if (object == null) return 2;
// 去除表头字符串中的空格
String objectStr = object.toString().replaceAll(" ", "");
if (rowString != null) rowString += "|" + objectStr;
else rowString = objectStr;
pcdataList.add(objectStr);
}
// 设置行节点
declList.add(new ElementDecl("row", "(" + rowString + ")*"));
// 遍历list设置行的下级节点
for (String tmp : pcdataList) {
declList.add(new ElementDecl(tmp, "(#PCDATA)"));
}
documentType.setInternalDeclarations(declList);
// 遍历读写excel数据到xml中
for (int x = 1; x < sheet.getLastRowNum(); x++) {
row = sheet.getRow(x);
Element rowElement = root.addElement("row");
for (int y = 0; y < row.getPhysicalNumberOfCells(); y++) {
// cell = row.getCell(y);
Object object = this.getCellValueObject(x, y);
if (object != null) {
// 将sheet第一行的行首元素当作元素名称
String pcdataString = pcdataList.get(y);
Element element = rowElement.addElement(pcdataString);
// Element element = rowElement.addElement("name");
element.setText(object.toString());
}
}
}
// 写入文件和dtd
document.setDocType(documentType);
this.docToXmlFile(document, xmlFile);
} catch (Exception e) {
e.printStackTrace();
}
return 1;
}
private static void confirmEmptyRow(Sheet s, int rowIx) {
Row row = s.getRow(rowIx);
assertTrue(row == null || row.getPhysicalNumberOfCells() == 0);
}
public static void readExcelFile(String fileName) {
FileInputStream file;
try {
file = new FileInputStream(new File(fileName));
Workbook excelFile = new HSSFWorkbook(file);
Sheet sheet1 = excelFile.getSheetAt(0); // Data sheet
// Set just in case metadata is incomplete or malformed
classificationLocation =
sheet1.getRow(0).getPhysicalNumberOfCells()
- 1; // Minus one since classificationLocation includes 0 and getPhysicalNumberOfCells
// does not
Sheet sheet2 = excelFile.getSheetAt(1); // Metadata sheet
// Loop based on number of attribute names
for (int i = 0; i < sheet2.getRow(0).getPhysicalNumberOfCells(); i++) {
String[] metadata = new String[METADATASIZE];
// Construct metadata
Row currRow = sheet2.getRow(0); // This should be a row of names
metadata[0] = currRow.getCell(i).toString();
currRow = sheet2.getRow(1); // This should be a row of data types (discrete or continuous)
metadata[1] = currRow.getCell(i).toString();
currRow = sheet2.getRow(2); // This should say which one is the classifier
if (currRow.getCell(i) == null
|| currRow.getCell(i).getCellType() == Cell.CELL_TYPE_BLANK) {
metadata[2] = "attribute";
} else {
metadata[2] = "classifier";
classificationLocation = i;
}
metadataLL.add(metadata);
}
for (Row row : sheet1) {
String data[] = new String[row.getPhysicalNumberOfCells() - 1];
int offset =
0; // Used so that we can declare an array of the size of the attributes without the
// classification
for (Cell cell : row) {
int index = cell.getColumnIndex();
if (classificationLocation != index) {
data[index - offset] = cell.toString();
} else {
classificationsLL.add(cell.toString());
// Moved to generate training data so that we do not get possible classifications from
// unknown data since some denote unknown by saying ?
// //Check to see if we have seen it yet
//
// int occurrences = 0;
// for(int i = 0; i < classificationTypes.size(); i++)
// {
// if(classificationTypes.get(i).compareTo(cell.toString()) == 0)
// {
// occurrences++;
// }
// }
// if(occurrences == 0)
// {
// classificationTypes.add(cell.toString());
// }
offset++;
}
}
dataLL.add(data);
// classCount = temp.length;
}
excelFile.close();
} catch (FileNotFoundException e) {
System.out.println("Error file not found");
System.exit(0);
} catch (IOException e) {
System.out.println("Unable to read file, disk drive may be failing");
e.printStackTrace();
System.exit(0);
}
}
