grouping observations accordeing to string dates with empty cells

Dear all,

I have the following matrix

[1] 'PDL12' '' [ NaN]

    [1]    'PDL12'    'JF 2009'    [-1.0000e-004]

    [1]    'PDL12'    'MA 2009'    [     -0.0330]

    [1]    'PDL12'    'MJ 2009'    [     -0.0268]

    [1]    'PDL12'    'JA 2009'    [     -0.0264]

    [1]    'PDL12'    'SO 2009'    [     -0.1913]

    [1]    'PDL12'    'ND 2009'    [     -0.0415]

    [1]    'PDL12'    'JF 2010'    [     -0.0168]

    [1]    'PDL12'    'MA 2010'    [     -0.0502]

    [1]    'PDL12'    'MJ 2010'    [     -0.0761]

    [1]    'PDL12'    'JA 2010'    [     -0.0429]

    [1]    'PDL12'    'SO 2010'    [     -0.0548]

    [1]    'PDL12'    'ND 2010'    [     -0.0173]

    [1]    'PDL12'    'JF 2011'    [      0.0071]
    [1]    'PDL12'    'MA 2011'    [      0.0121]

    [1]    'PDL12'    'MJ 2011'    [     -0.0727]

    [1]    'PDL12'    'JA 2011'    [     -0.1628]

    [1]    'PDL12'    'SO 2011'    [      0.0056]

    [2]    'PDL13'           ''    [         NaN]

    [2]    'PDL13'    'JF 2009'    [      0.0218]

    [2]    'PDL13'    'MA 2009'    [     -0.0037]

    [2]    'PDL13'    'MJ 2009'    [      0.0120]

    [2]    'PDL13'    'JA 2009'    [     -0.0027]

    [2]    'PDL13'    'SO 2009'    [-5.0000e-004]

    [2]    'PDL13'    'ND 2009'    [     -0.2818]

    [2]    'PDL13'    'JF 2010'    [     -0.0216]

    [2]    'PDL13'    'MA 2010'    [     -0.0330]

    [2]    'PDL13'    'MJ 2010'    [     -0.0468] 

    [2]    'PDL13'    'JA 2010'    [     -0.1309]

    [2]    'PDL13'    'SO 2010'    [     -0.0318]

    [2]    'PDL13'    'ND 2010'    [     -0.0014]

    [2]    'PDL13'    'JF 2011'    [     -0.0779]

    [2]    'PDL13'    'MA 2011'    [      0.0224]

    [2]    'PDL13'    'MJ 2011'    [     -0.0662]

    [2]    'PDL13'    'JA 2011'    [     -0.2187]

    [2]    'PDL13'    'SO 2011'    [     -0.0745]

    [3]    'PDL14'           ''    [         NaN]

    [3]    'PDL14'    'JF 2009'    [     -0.0107]

As you can see the date before JF2009 is empty. I would like to group all the observations that correspond to the empty cell, then all the observations that correspond to JF2009, then all the observations that correspond to MA2009 and similarly for the years 2011 and 2011

thanks

Answer 1

[Andrei Bobrov]

try

ms = cellstr(reshape(datestr(datenum(0,1:12,1,0,0,0),'m'),2,[])');
dd = regexprep(data(:,3),ms,cellstr(('1':num2str(numel(ms)))'));
dd(cellfun('isempty',dd)) = {'0 0000'};
[id,id] = sort(datenum(dd,'mm yyyy'));
out = data(id,:);

[antonet]

thank you andrei. it worked!