When saving a sheet of Calc document in .dbf format, the first row of the used range (R1) is parsed.
If the values of all R1 cells can be interpreted as Dbase field names (in particular, they are not empty), then they are interpreted as field names, the remaining rows are interpreted as database records. The field type is selected based on the data in the second row of the used range.
Otherwise, “N1”, “N2”, “N3”… are selected for field names, the field type is “Character”. All strings, including R1, are interpreted as database records. This is probably your case.
When opening a dBase file, Calc shows the field name, field type, length, and other information (depending on the field type) in the first line.
If necessary, you can upload your .ods file (only the first few rows are important) and we can analyze your case more specifically.
When you convert dBase to spreadsheet, you may ignore the extra descriptors or remove them with find&replace.
When opening a dBase database table in a spreadsheet (a spreadsheet application is NOT a database), Calc will always display the column types.
Just a heads-up that when you alter or remove that first row and save the .dbf file again, you may alter the database structure in names, types and lengths of fields. Which for most cases may not matter but routines relying on the structure may trip over it.
If you open a dBase-File in Calc you are allowed to make changes as you like, but when saving back Calc will tell you for example you can’t save more than 10 characters in the field Surname. If you use a numeric type for the first column you would be forced to put numeric values there or they can’t be saved.
.
So for dBase you can’t get rid of this. You could perhaps use another database, but then you can not open the file in Calc but would be forced to edit via Base or other software.
.
So the basic question is: Why did you choose dBase?
I need a dbf file so that I use it with an (old version) of Avery label pro, which doesn’t work with xlsx files. I know I should be using a later version, but that isn’t as good IMO.