A csv import will work a lot better if you structure your spreadsheet columns to match the TB address book. Create columns for First Name, Last Name, Display Name, Nickname, Primary Email, Secondary Email, Screen Name, Work Phone, Home Phone, Fax Number, Pager Number, Mobile Number (and the rest). Leave a column empty if you don't have any data. Check that the input fields match the address book fields and move them up/down in the import dialog.

Yeah, did all that. In the import dialogue box, clicking on "Next" shows the correct fields and data and I can step through the entire 107 entries and all show as correct. But clicking on "Finish" to create the file gives an Address Book file with only two fields, the Name one and the email one. Name is filled with the beginning of each email address up to the @ symbol, the second field doesn't exist and the email field has the correct email addresses.

Without being able to see the csv, it's hard to explain your result. If you view the csv in a text editor, does it look correct, e.g. there are commas and not colons? If you create a test address book in TB with a few contacts, then export it to a csv (Tools/Export, Save as type:csv), does it look similar to the csv you exported from the spreadsheet?

Opening the .csv file in Notepad, it looks exactly right. The three fields are delimited by commas. And single stepping back and forth through the import process with the "Next" and "Previous" buttons shows the data in the three fields as specified. But creating the file by clicking on "OK" gives the wrong result. Exporting a small test address book as a .csv shows the same data separated by commas, just as my import file. Except the field names are different, but that is to be expected.

try saving the file in notepad using save as and specifying Unicode in the save as dialog.

Excel in particular writes ANSI text which the importer does not understand as it is basically obsolete except in Microsoft Office.

Tried switching from ANSI to Unicode and it made no difference. This is very puzzling!

Further to that, when I select an entry in my imported address book and click "edit", it shows the correct data in the edit panel. So it appears to be a problem with the way Thunderbird displays the imported address book data rather than actually importing the data.

I have solved it now, I needed to match up "Display Name" to the name data in my .csv file. In the Address Book display in Thunderbird, it has a greyed out field called "Name" which is unchangeable and appears first in the listing. This is actually the Display Name field although it doesn't say so. The import process has worked perfectly all along, its just the way the fields are named that is the problem. What Geeky software!