Corrupted output file when adding a table with "header_row: True" and no data

I am using XlsxWriter to create tables with a header row and data underneath. Sometimes my script has no data so there is only one row which is the header row. In that case, the generated file is corrupted according to Excel.

I am using Python version 3.7.3 and XlsxWriter 1.2.8 and Excel version 2016.

Here is some code that demonstrates the problem:

#!/usr/bin/env python3
import xlsxwriter

workbook = xlsxwriter.Workbook("test.xlsx")
worksheet = workbook.add_worksheet()

data = []

worksheet.add_table(first_row=0, first_col=0,
                    last_row=len(data), last_col=1,
                    options={
                        'columns': [{'header': 'test1'}, {'header': 'test2'}],
                        'data': data,
                        'header_row': True
                    })

workbook.close()

I know that the error is on my side but IMO xlsxwriter should catch such case and trigger an error.

Actually here is how Excel handles this:

1. Create new file
2. A1=“test 1” and B1="test 2
3. Select A1 and B1
4. Insert table and check “my table as headers”
5. Notice that the table now covers two rows (so one empty data row was added)

So xlsxwriter could try to fix it in the same spirit, for example:

if len(data)==0 and last_row==0 and options['header_row']:
    last_row=1

Which adds an empty data row and generates a valid file.

[jmcnamara]

Fixed on master. The module now issues a warning for tables without space for data rows. The header row, or lack of one, is taken into account. Thanks for the report.

[jmcnamara]

The other one is slightly different I think.

Ok. Thanks. I’ll handle them separately.

Maybe both could be fixed by adding one empty row when there’s no data row.

In general XlxsWriter doesn’t try to fix input issue for the user. This will probably be fixed with a warning.