A Byte order mark (BOM) inserted by writetable (xml export and xlsx export), why?

A byte order mark (BOM) is inserted by writetable when exporting a table to a CSV file, using this command:

writetable(T, ‘T.csv’, 'Encoding', 'UTF-8')

This zero-length BOM character (U+FEFF) is seen in the following image taken from a text editor (showing invisibles), the BOM being pictured as a red dot (invisible character) at the very first position of the file content.

I can understand the logic behind inserting this BOM (helping read a text file with a specific encoding), even if I prefer the BOM addition to be optional.

What I don’t understand is the rationale of inserting a BOM in each tag of the same table, exported in XML format (see the red dots before each of the UNIX_TIMESTAMP tags).

It doesn't make sense to me. The UNIX_TIMESTAMP tags should not have a BOM, right? Or is there a particular reason?

PS: what would be the most efficient way to get rid of the BOM(s)?

PS2: the BOMs are also present in the UNIX_TIMESTAMP cell (cell A1) when exporting to XLSX. This is not right too. It should only contain a string with ASCII characters in the current case.

Many thanks

[Walter Roberson]

Huh. Strange.

[Stephane]

I digged further, and the problem does not come from writetable, my bad. Will prepare an answer.

Answer 1

[Stephane]

The problem does not come from writetable actually, but from a hidden character in the char values contained in the cell array defining VariableNames.

... 3 hours on this ... but at least I solved it :0

'UNIX_TIMESTAMP' contained an invisible character at the start of the char.

t_equal_cell = table(timestamp, x_g, VariableNames={'UNIX_TIMESTAMP','Gx'});

[Stephane]

I will leave this answer there in case it helps.