Don't put spaces in CSV output like that, it may very well break (some) readers/parsers.

If your CSV - as in your example - only has numbers, be aware that if the current locale has "," as TRIAD_SEP you can get into weird problems (maybe not in the perl part, but you are talking about a tool chain here). If you want to make sure your CSV is valid CSV, use a module like Text::CSV and/or Text::CSV_XS to generate the CSV.

I have not tested with recent Excel versions, but what bitimgduck suggests, might work, but then again, might not because Excel is known to ignore "strings" in CSV imports. If the value after parsing looks like a date (M$Excel speak), it will convert it to a date. Most of the time, it will do it wrong, as "date" in Excel is always US date format (MM/DD/YYYY), even if the current locale is close to ISO (YYYY-MM-DD or DD-MM-YYYY). If your column consists of plain "numbers" (not dates) and one of the "numbers" comes up as a date in the Excel sheet after import, do not be surprised.

Then there is three ways of opening a CSV in Excel:

• Double-click on the csv file in the explorer;
• Open Excele, and then Alt-F->Open the file;
• Open Excel and then Data->Import from file.

These all interpret CSV data differently! YMMV

So, the only way to make your data look consistent in Excel is to write an Excel book yourself, and - as you have shown already - use formatting specifications. Spreadsheet::WriteExcel works perfect and is able to write reproducible formatting.