[BUG] addRow() not adding merged cells correctly which are read from source file

Hi All,

I am trying to read from a excel file, manipulate some data and write to another excel. I am seeing a weird issue when adding rows which have merged cells. They are not getting added as expected in the destination file, the rows are appearing non-merged.

Below is the relevant part of the code. I am using streams and addRow.

const readerOptions = {
sharedStrings: 'cache',
 hyperlinks: 'cache',
 worksheets: 'emit',
 styles: 'cache',
};
const writerOptions = {
 filename: fpath,
 useStyles: true,
 useSharedStrings: true
}
const workbook = new ExcelJS.stream.xlsx.WorkbookWriter(writerOptions)
const myworksheet = workbook.addWorksheet('Sheet 1');
const workbookReader = new ExcelJS.stream.xlsx.WorkbookReader(sheet.path, readerOptions);
workbookReader.read();
workbookReader.on('worksheet', worksheet => {
 worksheet.on('row', row => {
  const r = myworksheet.addRow();
  Object.assign(r, row);
 });
});

workbookReader.on('end', async () => {
 await workbook.commit();
});

Please find the below images for your reference. first image is actual_result.png and second image is expected_result.png

Any help on how to fix this will be really great, Thanks.

[eug-vs]

Hi! Could this be related to #292? My use-case is duplicating rows: currently duplicating a row with merged cells results into a row with all cells unmerged. I’ve came up with a workaround (it assumes that row you’re duplicating only contains merged cells within itself, i.e no cell in this row is merged with other rows):

const duplicateRowWithMergedCells = (sheet: any, row: number, count: number) => {
  sheet.duplicateRow(row, count, true);

  const merges: string[] = sheet.model.merges;
  // Find all merges inside initial row
  const rowMerges = merges.filter(range => range.match(`\\w+${row}:\\w+${row}`));

  _.times(count, index => {
    const newRow = row + index + 1;

    // Unmerge everything in a newRow so we dont run into conflicts
    merges
      .filter(range => range.match(`\\w+${newRow}:\\w+${newRow}`))
      .forEach(range => sheet.unMergeCells(range));

    // Merge the same cells as in the initial row
    rowMerges
      .map(range => range.replace(new RegExp(`${row}`, 'g'), `${newRow}`))
      .forEach(range => sheet.mergeCells(range));
  });
};

Do you think the above way makes sense or may be there might be a better way to detect if rows has merged cells.

@aksharj note how I unmerge rows before manually transferring new merges. I think this answers you question.

I did not dive deep into the library, but my understanding of this part of the bug is: when duplicating a row with insert = true (and maybe in other scenarios) it does not remove old merges from cells which were previously on the positions where are you inserting. Therefore you can run into conflicts sometimes.

[aksharj]

hi @Siemienik thanks for your reply. Actually, excel files in some case might be large that’s the reason i need to use streams.

I have found a workaround, i am checking if the row has merged cells and then merge it again after adding rows. The way i am checking if rows contain merged cells is as follows.

if((newRow.actualCellCount !== newRow.cellCount) && (newRow.actualCellCount == 1)) {

Do you think the above way makes sense or may be there might be a better way to detect if rows has merged cells.

Full relevant code to copy excel is as follows.

        const workbook = new ExcelJS.stream.xlsx.WorkbookWriter(writerOptions);
        const workbookReader = new ExcelJS.stream.xlsx.WorkbookReader(sheet.path, readerOptions);
        const myworksheet = workbook.addWorksheet('Sheet 1');
        workbookReader.read();
        workbookReader.on('worksheet', worksheet => {
          worksheet.on('row', row => {
            const newRow = myworksheet.addRow();
            Object.assign(newRow, row);
            if((newRow.actualCellCount !== newRow.cellCount) && (newRow.actualCellCount == 1)) {
              let cellLength = newRow.model.cells.length;
              myworksheet.mergeCells(`${newRow.model.cells[0].address}:${newRow.model.cells[cellLength - 1].address}`)
            }
          });
        });

Do let me know if this makes sense or may be better way to achieve this, for now this works for me.

Thanks.