How to plot data with corresponding dates

I have a vector with data (1x9324 double) and another vector (1x9324 cell) with the corresponding dates. Both vectors are from an Excel sheet using 'xlsread'. The dates are in string format and not equidistant. e.g.

    '02.01.2007 08:45:00'
            to
    '31.12.2008 12:27:00'
    '31.12.2008 15:31:00'
    '31.12.2008 16:35:00'
    '31.12.2008 16:35:00'
    '31.12.2008 16:35:00'

There are also dates wich are exactly the same. But I do not want to interpolate or take averages. I want to plot the data on the Y axis and the dates on the X axis, just month and year are enough. I tryed

    startDate = datenum('02.01.2007');
    endDate = datenum('31.12.2008');
    xData = linspace(startDate,endDate,9324);
    plot(xData,data_vector);
    datetick('x','mmm,yy','keepticks');

But I'm afreid that the dates are now equidistant. Then I tryed

ts = timeseries (data_vector, datenum(date_vector));
plot(ts);

But both vectors changed in length and values. Maybe because the timeseries tool is taking averages of the data where the dates are the same.

System:

MATLAB 7.10.0 (R2010a)

Microsoft Excel 97-2003 Worksheet (.xls)

Windows 8.1 64Bit

Thanks in advance!

Answer 1

[David Young]

The problem is that you are using linspace to produce equally separated dates rather than converting each of the actual dates. Try

plot(datenum(dates, 'dd.mm.yyyy HH:MM:SS'), data_vector);

where dates is the cell array of strings containing your dates.

[Answer edited to include format string for datenum]

[Karl-Martin]

When I do the following

    plot(datenum(dates), data_vector);
    datetick('x','mmm,yy','keepticks');

I get this graph

It seems like the x values are not clearly linked to the y values and the dates look all the same. Maybe there is something wrong with the date conversion.

When I test my first date with

    datenum('02.01.2007 08:45:00')
    ans = 7.3560e+005

And back again with

    datestr(ans)

I get a different date

    ans =

    01-Jan-2014 08:45:00

[David Young]

Yes, you're right - the date string isn't in the standard format, so isn't being converted correctly. This should work though:

datenum(dates, 'dd.mm.yyyy HH:MM:SS')

[David Young]

Note: answer edited to reflect this.