haridsv commented on code in PR #1794: URL: https://github.com/apache/phoenix/pull/1794#discussion_r1453723232
########## phoenix-core/src/main/java/org/apache/phoenix/coprocessor/CDCGlobalIndexRegionScanner.java: ########## @@ -104,115 +94,74 @@ protected Scan prepareDataTableScan(Collection<byte[]> dataRowKeys) throws IOExc protected boolean getNextCoveredIndexRow(List<Cell> result) throws IOException { if (indexRowIterator.hasNext()) { List<Cell> indexRow = indexRowIterator.next(); - for (Cell c: indexRow) { - if (c.getType() == Cell.Type.Put) { - result.add(c); - } - } + Cell firstCell = indexRow.get(indexRow.size() - 1); + byte[] indexRowKey = new ImmutableBytesPtr(firstCell.getRowArray(), + firstCell.getRowOffset(), firstCell.getRowLength()) + .copyBytesIfNecessary(); + ImmutableBytesPtr dataRowKey = new ImmutableBytesPtr( + indexToDataRowKeyMap.get(indexRowKey)); + Result dataRow = dataRows.get(dataRowKey); + Long indexCellTs = firstCell.getTimestamp(); + Cell.Type indexCellType = firstCell.getType(); + + Map<ImmutableBytesPtr, Cell> preImageObj = new HashMap<>(); + Map<ImmutableBytesPtr, Cell> changeImageObj = new HashMap<>(); + List<Cell> resultCells = Arrays.asList(dataRow.rawCells()); + Collections.sort(resultCells, CellComparator.getInstance().reversed()); + + boolean isIndexCellDeleteRow = false; + boolean isIndexCellDeleteColumn = false; try { - Result dataRow = null; - if (! result.isEmpty()) { - Cell firstCell = result.get(0); - byte[] indexRowKey = new ImmutableBytesPtr(firstCell.getRowArray(), - firstCell.getRowOffset(), firstCell.getRowLength()) - .copyBytesIfNecessary(); - ImmutableBytesPtr dataRowKey = new ImmutableBytesPtr( - indexToDataRowKeyMap.get(indexRowKey)); - dataRow = dataRows.get(dataRowKey); - Long indexRowTs = result.get(0).getTimestamp(); - Map<Long, Map<ImmutableBytesPtr, Cell>> changeTimeline = dataRowChanges.get( - dataRowKey); - if (changeTimeline == null) { - List<Cell> resultCells = Arrays.asList(dataRow.rawCells()); - Collections.sort(resultCells, CellComparator.getInstance().reversed()); - List<Cell> deleteMarkers = new ArrayList<>(); - List<List<Cell>> columns = new LinkedList<>(); - Cell currentColumnCell = null; - Pair<byte[], byte[]> emptyKV = EncodedColumnsUtil.getEmptyKeyValueInfo( - EncodedColumnsUtil.getQualifierEncodingScheme(scan)); - List<Cell> currentColumn = null; - Set<Long> uniqueTimeStamps = new HashSet<>(); - // TODO: From CompactionScanner.formColumns(), see if this can be refactored. - for (Cell cell : resultCells) { - uniqueTimeStamps.add(cell.getTimestamp()); - if (cell.getType() != Cell.Type.Put) { - deleteMarkers.add(cell); - } - if (CellUtil.matchingColumn(cell, QueryConstants.DEFAULT_COLUMN_FAMILY_BYTES, - emptyKV.getFirst())) { - continue; - } - if (currentColumnCell == null) { - currentColumn = new LinkedList<>(); - currentColumnCell = cell; - currentColumn.add(cell); - } else if (!CellUtil.matchingColumn(cell, currentColumnCell)) { - columns.add(currentColumn); - currentColumn = new LinkedList<>(); - currentColumnCell = cell; - currentColumn.add(cell); - } else { - currentColumn.add(cell); - } + for (Cell cell : resultCells) { + if (cell.getType() == Cell.Type.DeleteColumn) { + // DDL is not supported in CDC Review Comment: Also, there can be multiple columns set to NULL in the same UPSERT statements, so we would have to detect and surface these changes as is. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@phoenix.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org