Where did this CSV file originate?
The file originates as a clipboard-copied table from this online article (paywall, sorry). I then converted the table via LibreOffice to a sheet and saved this as a CSV.
Thanks for the detailed explanation @NickLowe. This is soooooo cool.
It think this would be worth to be mentioned from official side – or is the feature to “pin” the first row as header via #text suffixes not here to stay, @cgrunenberg?
Perhaps it’s a product of the conversion but it’s certainly not a well-formed .csv file 
I think this would be worth to be mentioned from official side – or is the feature to “pin” the first row as header via
#textsuffixes not here to stay, @cgrunenberg?
Since DEVONthink 3 came out, sheets have always been created with the data type set in the text, but hidden from the user. And #text is not the only data type.
Some more context: this table hails from a print magazine originally. The first row (bold) shows the product names, the first column is used to display the feature names. Not so unusual for a print magazine: readers can thus compare the features of products side by side. Pivoting the table would make that more difficult.
So, the logical “header” is in the first column, not the first row. Which is not what a spreadsheet expects from a CSV. Basically, the OP got exactly what they input – minus the bold font in the first row.
Yes, but that works only, when I tell DT via the #text hack, that the first row is really the header.
I was talking about the original table at heise.de here, since not everyone can see that.