How to modify a large Excel file when memory is an issue

As told in comments above there is no one fits all solution using pure XML reading and writing the Office Open XML spreadsheets. Each Excel workbook needs it's own code dependent on it's structure and on what content shall be changed.

This is because apache poi's high level classes provides a meta level to avoid this. But this needs memory to work. And for very big workbooks it needs much memory. To avoid memory consumption through manipulating the XML directly this meta level is not usable. So one must know the XML structure of a worksheet and the meaning of the XML elements used.

So if we have a Excel workbook having a first sheet having strings in column A and numbers in column B, then we could changing every fifth row using StAX for manipulating the XML directly using following code:

import org.apache.poi.openxml4j.opc.OPCPackage;
import org.apache.poi.openxml4j.opc.PackagePart;

import org.apache.poi.xssf.model.SharedStringsTable;
import org.apache.poi.xssf.usermodel.XSSFRichTextString;

import org.openxmlformats.schemas.spreadsheetml.x2006.main.CTRst;

import javax.xml.stream.XMLEventFactory;
import javax.xml.stream.XMLEventReader;
import javax.xml.stream.XMLEventWriter;
import javax.xml.stream.XMLInputFactory;
import javax.xml.stream.XMLOutputFactory;
import javax.xml.stream.events.Characters;
import javax.xml.stream.events.StartElement;
import javax.xml.stream.events.XMLEvent;

import javax.xml.namespace.QName;

import java.io.File;
import java.io.InputStream;
import java.io.OutputStream;

import java.util.regex.Pattern;

class StaxReadAndChangeTest {

 public static void main(String[] args) throws Exception {
  File file = new File("ReadAndWriteTest.xlsx");
  OPCPackage opcpackage = OPCPackage.open(file);

  //since there are strings in the sheet data, we need the SharedStringsTable
  PackagePart sharedstringstablepart = opcpackage.getPartsByName(Pattern.compile("/xl/sharedStrings.xml")).get(0);
  SharedStringsTable sharedstringstable = new SharedStringsTable();
  sharedstringstable.readFrom(sharedstringstablepart.getInputStream());

  //get first worksheet
  PackagePart sheetpart = opcpackage.getPartsByName(Pattern.compile("/xl/worksheets/sheet1.xml")).get(0);

  //get XML reader and writer            
  XMLEventReader reader = XMLInputFactory.newInstance().createXMLEventReader(sheetpart.getInputStream());
  XMLEventWriter writer = XMLOutputFactory.newInstance().createXMLEventWriter(sheetpart.getOutputStream());

  XMLEventFactory eventFactory = XMLEventFactory.newInstance();

  int rowsCount = 0;
  int colsCount = 0;
  boolean cellAfound = false;
  boolean cellBfound = false;

  while(reader.hasNext()){ //loop over all XML in sheet1.xml
   XMLEvent event = (XMLEvent)reader.next();
   if(event.isStartElement()) {
    StartElement startElement = (StartElement)event;
    QName startElementName = startElement.getName();
    if(startElementName.getLocalPart().equalsIgnoreCase("row")) { //start element of row
     rowsCount++;
     colsCount = 0;
    } else if (startElementName.getLocalPart().equalsIgnoreCase("c")) { //start element of cell
     colsCount++;
     cellAfound = false;
     cellBfound = false;
     if (rowsCount % 5 == 0) { // every 5th row
      if (colsCount == 1) { // cell A
       cellAfound = true;
      } else if (colsCount == 2) { // cell B
       cellBfound = true;
      } 
     }
    } else if (startElementName.getLocalPart().equalsIgnoreCase("v")) { //start element of value
     if (cellAfound) {
      // create new rich text content for cell A
      CTRst ctstr = CTRst.Factory.newInstance();
      ctstr.setT("changed String Value A" + (rowsCount));
      //int sRef = sharedstringstable.addEntry(ctstr);
      int sRef = sharedstringstable.addSharedStringItem(new XSSFRichTextString(ctstr));
      // set the new characters for A's value in the XML
      if (reader.hasNext()) {
       writer.add(event); // write the old event
       event = (XMLEvent)reader.next(); // get next event - should be characters
       if (event.isCharacters()) {
        Characters value = eventFactory.createCharacters(Integer.toString(sRef));
        event = value;
       } 
      }        
     } else if (cellBfound) {
      // set the new characters for B's value in the XML
      if (reader.hasNext()) {
       writer.add(event); // write the old event
       event = (XMLEvent)reader.next(); // get next event - should be characters
       if(event.isCharacters()) { 
        double oldValue = Double.valueOf(((Characters)event).getData()); // old double value
        Characters value = eventFactory.createCharacters(Double.toString(oldValue * rowsCount));
        event = value;         
       }         
      }
     }
    }
   }
   writer.add(event); //by default write each read event
  }
  writer.flush();

  //write the SharedStringsTable
  OutputStream out = sharedstringstablepart.getOutputStream();
  sharedstringstable.writeTo(out);
  out.close();
  opcpackage.close();

 }
}

This will be much less memory consuming than apache poi's XSSF classes. But, as said, it only works exactly for this kind of Excel workbook having a first sheet having strings in column A and numbers in column B.