[Gnuplot-info] missing data = empty string ?

[Theo <th...@gm...>]

Hi there,
I have data files with missing data.
e.g. the following TAB-separated file with 4 columns.

13.05.2013	0	1	16.4
14.05.2013	0		16.5
15.05.2013	0		16.2
16.05.2013	0		17
17.05.2013	0		16.6
18.05.2013	0	1	
19.05.2013	0	1	
20.05.2013	0	1	17.2
21.05.2013	0		16.8
22.05.2013	0		16.9
23.05.2013	0		16.9

If I want to plot column 4 gnuplot seems to plot only the data where 
there are 4 values in the row. If I want to plot column 3 it plots 
mostly data of column 4.
Since gnuplot has the option to define how it should handle missing data
I expected that

   set datafile missing ""    # empty string

will solve the issue. But it doesn't work.
   plot 'Data.dat' u 1:4 with lines
   plot 'Data.dat' u 1:($4) with lines
doesn't work either.
And 'help missing' doesn't help me further.
I do not want to change all the missing data to '?' or 'NaN'.
Any spreadsheet program manages to import correctly such data with empty 
fields. So I expected that gnuplot can also do it...
Any option that I missed?
Thanks, Theo.