Avoid spurious missing lines with the MissingRecordAware event code, and odd files that contain RowRecords in the middle of the cell Records.

git-svn-id: https://svn.apache.org/repos/asf/poi/trunk@672562 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Nick Burch 2008-06-28 18:21:21 +00:00
parent 1e0458db99
commit 70f9a0a528
5 changed files with 112 additions and 35 deletions

View File

@ -37,6 +37,7 @@
<!-- Don't forget to update status.xml too! -->
<release version="3.1.1-alpha1" date="2008-??-??">
<action dev="POI-DEVELOPERS" type="fix">Avoid spurious missing lines with the MissingRecordAware event code, and odd files that contain RowRecords in the middle of the cell Records.</action>
<action dev="POI-DEVELOPERS" type="add">Support for parsing formulas during EventUserModel processing, via the new EventWorkbookBuilder</action>
</release>
<release version="3.1-final" date="2008-06-29">

View File

@ -34,6 +34,7 @@
<!-- Don't forget to update changes.xml too! -->
<changes>
<release version="3.1.1-alpha1" date="2008-??-??">
<action dev="POI-DEVELOPERS" type="fix">Avoid spurious missing lines with the MissingRecordAware event code, and odd files that contain RowRecords in the middle of the cell Records.</action>
<action dev="POI-DEVELOPERS" type="add">Support for parsing formulas during EventUserModel processing, via the new EventWorkbookBuilder</action>
</release>
<release version="3.1-final" date="2008-06-29">

View File

@ -46,8 +46,14 @@ import org.apache.poi.hssf.record.RowRecord;
*/
public class MissingRecordAwareHSSFListener implements HSSFListener {
private HSSFListener childListener;
private int lastSeenRow = -1;
private int lastSeenColumn = -1;
// Need to have different counters for cell rows and
// row rows, as you sometimes get a RowRecord in the
// middle of some cells, and that'd break everything
private int lastRowRow = -1;
private int lastCellRow = -1;
private int lastCellColumn = -1;
/**
* Constructs a new MissingRecordAwareHSSFListener, which
@ -71,14 +77,16 @@ public class MissingRecordAwareHSSFListener implements HSSFListener {
if (bof.getType() == bof.TYPE_WORKBOOK)
{
// Reset the row and column counts - new workbook
lastSeenRow = -1;
lastSeenColumn = -1;
lastRowRow = -1;
lastCellRow = -1;
lastCellColumn = -1;
//System.out.println("Encountered workbook");
} else if (bof.getType() == bof.TYPE_WORKSHEET)
{
// Reset the row and column counts - new sheet
lastSeenRow = -1;
lastSeenColumn = -1;
lastRowRow = -1;
lastCellRow = -1;
lastCellColumn = -1;
//System.out.println("Encountered sheet reference");
}
break;
@ -92,15 +100,15 @@ public class MissingRecordAwareHSSFListener implements HSSFListener {
// + rowrec.getFirstCol() + " last column at " + rowrec.getLastCol());
// If there's a jump in rows, fire off missing row records
if(lastSeenRow+1 < rowrec.getRowNumber()) {
for(int i=(lastSeenRow+1); i<rowrec.getRowNumber(); i++) {
if(lastRowRow+1 < rowrec.getRowNumber()) {
for(int i=(lastRowRow+1); i<rowrec.getRowNumber(); i++) {
MissingRowDummyRecord dr = new MissingRowDummyRecord(i);
childListener.processRecord(dr);
}
}
// Record this as the last row we saw
lastSeenRow = rowrec.getRowNumber();
lastRowRow = rowrec.getRowNumber();
break;
@ -157,45 +165,49 @@ public class MissingRecordAwareHSSFListener implements HSSFListener {
break;
}
// Do we need to fire dummy end-of-row records?
if(thisRow != lastSeenRow) {
for(int i=lastSeenRow; i<thisRow; i++) {
// If we're on cells, and this cell isn't in the same
// row as the last one, then fire the
// dummy end-of-row records?
if(thisRow != lastCellRow && lastCellRow > -1) {
for(int i=lastCellRow; i<thisRow; i++) {
int cols = -1;
if(i == lastSeenRow) {
cols = lastSeenColumn;
if(i == lastCellRow) {
cols = lastCellColumn;
}
LastCellOfRowDummyRecord r = new LastCellOfRowDummyRecord(i, cols);
childListener.processRecord(r);
}
}
// If we've finished with the columns, then fire the final
// dummy end-of-row record
if(lastSeenRow != -1 && lastSeenColumn != -1 && thisRow == -1) {
LastCellOfRowDummyRecord r = new LastCellOfRowDummyRecord(lastSeenRow, lastSeenColumn);
// If we've just finished with the cells, then fire the
// final dummy end-of-row record
if(lastCellRow != -1 && lastCellColumn != -1 && thisRow == -1) {
LastCellOfRowDummyRecord r = new LastCellOfRowDummyRecord(lastCellRow, lastCellColumn);
childListener.processRecord(r);
lastSeenRow = -1;
lastSeenColumn = -1;
lastCellRow = -1;
lastCellColumn = -1;
}
// If we've moved onto a new row, the ensure we re-set
// the column counter
if(thisRow != lastSeenRow) {
lastSeenColumn = -1;
if(thisRow != lastCellRow) {
lastCellColumn = -1;
}
// Do we need to fire dummy cell records?
if(lastSeenColumn != (thisColumn-1)) {
for(int i=lastSeenColumn+1; i<thisColumn; i++) {
// If there's a gap in the cells, then fire
// the dummy cell records?
if(lastCellColumn != (thisColumn-1)) {
for(int i=lastCellColumn+1; i<thisColumn; i++) {
MissingCellDummyRecord r = new MissingCellDummyRecord(thisRow, i);
childListener.processRecord(r);
}
}
// Update cell and row counts if doing cells
// Update cell and row counts as needed
if(thisColumn != -1) {
lastSeenRow = thisRow;
lastSeenColumn = thisColumn;
lastCellColumn = thisColumn;
lastCellRow = thisRow;
}
childListener.processRecord(record);

View File

@ -16,8 +16,6 @@
==================================================================== */
package org.apache.poi.hssf.eventusermodel;
import java.io.File;
import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.util.ArrayList;
@ -29,6 +27,7 @@ import org.apache.poi.hssf.HSSFTestDataSamples;
import org.apache.poi.hssf.eventusermodel.dummyrecord.LastCellOfRowDummyRecord;
import org.apache.poi.hssf.eventusermodel.dummyrecord.MissingCellDummyRecord;
import org.apache.poi.hssf.eventusermodel.dummyrecord.MissingRowDummyRecord;
import org.apache.poi.hssf.record.BOFRecord;
import org.apache.poi.hssf.record.LabelSSTRecord;
import org.apache.poi.hssf.record.Record;
import org.apache.poi.hssf.record.RowRecord;
@ -40,8 +39,7 @@ public final class TestMissingRecordAwareHSSFListener extends TestCase {
private Record[] r;
public void setUp() {
public void openNormal() {
HSSFRequest req = new HSSFRequest();
MockHSSFListener mockListen = new MockHSSFListener();
MissingRecordAwareHSSFListener listener = new MissingRecordAwareHSSFListener(mockListen);
@ -55,10 +53,31 @@ public final class TestMissingRecordAwareHSSFListener extends TestCase {
} catch (IOException e) {
throw new RuntimeException(e);
}
r = mockListen.getRecords();
assertTrue(r.length > 100);
}
public void openAlt() {
HSSFRequest req = new HSSFRequest();
MockHSSFListener mockListen = new MockHSSFListener();
MissingRecordAwareHSSFListener listener = new MissingRecordAwareHSSFListener(mockListen);
req.addListenerForAllRecords(listener);
HSSFEventFactory factory = new HSSFEventFactory();
try {
InputStream is = HSSFTestDataSamples.openSampleFileStream("MRExtraLines.xls");
POIFSFileSystem fs = new POIFSFileSystem(is);
factory.processWorkbookEvents(req, fs);
} catch (IOException e) {
throw new RuntimeException(e);
}
r = mockListen.getRecords();
assertTrue(r.length > 100);
}
public void testMissingRowRecords() throws Exception {
openNormal();
// We have rows 0, 1, 2, 20 and 21
int row0 = -1;
@ -108,6 +127,7 @@ public final class TestMissingRecordAwareHSSFListener extends TestCase {
}
public void testEndOfRowRecords() throws Exception {
openNormal();
// Find the cell at 0,0
int cell00 = -1;
@ -194,7 +214,7 @@ public final class TestMissingRecordAwareHSSFListener extends TestCase {
assertTrue(r[cell00+57] instanceof LastCellOfRowDummyRecord);
// Check the numbers of the last seen columns
LastCellOfRowDummyRecord[] lrs = new LastCellOfRowDummyRecord[23];
LastCellOfRowDummyRecord[] lrs = new LastCellOfRowDummyRecord[24];
int lrscount = 0;
for(int i=0; i<r.length; i++) {
if(r[i] instanceof LastCellOfRowDummyRecord) {
@ -229,6 +249,7 @@ public final class TestMissingRecordAwareHSSFListener extends TestCase {
public void testMissingCellRecords() throws Exception {
openNormal();
// Find the cell at 0,0
int cell00 = -1;
@ -326,10 +347,41 @@ public final class TestMissingRecordAwareHSSFListener extends TestCase {
assertEquals(22, mc.getRow());
assertEquals(10, mc.getColumn());
}
// Make sure we don't put in any extra new lines
// that aren't already there
public void testNoExtraNewLines() throws Exception {
// Load a different file
openAlt();
for(int i=0; i<r.length; i++) {
System.err.println(r[i]);
}
// This file has has something in lines 1-33
List lcor = new ArrayList();
for(int i=0; i<r.length; i++) {
if(r[i] instanceof LastCellOfRowDummyRecord)
lcor.add( (LastCellOfRowDummyRecord)r[i] );
}
// Check we got the 33 rows
assertEquals(33, lcor.size());
LastCellOfRowDummyRecord[] rowEnds = (LastCellOfRowDummyRecord[])
lcor.toArray(new LastCellOfRowDummyRecord[lcor.size()]);
assertEquals(33, rowEnds.length);
// And check they have the right stuff in them,
// no repeats
for(int i=0; i<rowEnds.length; i++) {
assertEquals(i, rowEnds[i].getRow());
}
}
private static final class MockHSSFListener implements HSSFListener {
public MockHSSFListener() {}
private final List _records = new ArrayList();
private boolean logToStdOut = false;
public void processRecord(Record record) {
_records.add(record);
@ -346,9 +398,20 @@ public final class TestMissingRecordAwareHSSFListener extends TestCase {
LastCellOfRowDummyRecord lc = (LastCellOfRowDummyRecord)record;
log("Got end-of row, row was " + lc.getRow() + ", last column was " + lc.getLastColumnNumber());
}
if(record instanceof BOFRecord) {
BOFRecord r = (BOFRecord)record;
if(r.getType() == BOFRecord.TYPE_WORKSHEET) {
log("On new sheet");
}
}
if(record instanceof RowRecord) {
RowRecord rr = (RowRecord)record;
log("Starting row #" + rr.getRowNumber());
}
}
private static void log(String msg) {
if(false) { // successful tests should be quiet
private void log(String msg) {
if(logToStdOut) {
System.out.println(msg);
}
}