gathering together specific elements from a column vector

HI all, I have the matrix

A=

identif pairs time allpairs

 [1]    'PDL12'    'JF2008'    [         NaN]
    [1]    'PDL12'    'JF2009'    [-1.0000e-004]
    [1]    'PDL12'    'MA2009'    [     -0.0330]
    [1]    'PDL12'    'MJ2009'    [     -0.0268]
    [1]    'PDL12'    'JA2009'    [     -0.0264]
    [1]    'PDL12'    'SO2009'    [     -0.1913]
    [1]    'PDL12'    'ND2009'    [     -0.0415]
    [1]    'PDL12'    'JF2010'    [     -0.0168]
    [1]    'PDL12'    'MA2010'    [     -0.0502]
    [1]    'PDL12'    'MJ2010'    [     -0.0761]
    [1]    'PDL12'    'JA2010'    [     -0.0429]
    [1]    'PDL12'    'SO2010'    [     -0.0548]
    [1]    'PDL12'    'ND2010'    [     -0.0173]
    [1]    'PDL12'    'JF2011'    [      0.0071]
    [1]    'PDL12'    'MA2011'    [      0.0121]
    [1]    'PDL12'    'MJ2011'    [     -0.0727]
    [1]    'PDL12'    'JA2011'    [     -0.1628]
    [1]    'PDL12'    'SO2011'    [      0.0056]
    [2]    'PDL13'    'JF2008'    [         NaN]
    [2]    'PDL13'    'JF2009'    [      0.0218]
    [2]    'PDL13'    'MA2009'    [     -0.0037]
    [2]    'PDL13'    'MJ2009'    [      0.0120]
    [2]    'PDL13'    'JA2009'    [     -0.0027]
    [2]    'PDL13'    'SO2009'    [-5.0000e-004]
    [2]    'PDL13'    'ND2009'    [     -0.2818]
    [2]    'PDL13'    'JF2010'    [     -0.0216]
    [2]    'PDL13'    'MA2010'    [     -0.0330]
    [2]    'PDL13'    'MJ2010'    [     -0.0468]
    [2]    'PDL13'    'JA2010'    [     -0.1309]
    [2]    'PDL13'    'SO2010'    [     -0.0318]
    [2]    'PDL13'    'ND2010'    [     -0.0014]
    [2]    'PDL13'    'JF2011'    [     -0.0779]
    [2]    'PDL13'    'MA2011'    [      0.0224]
    [2]    'PDL13'    'MJ2011'    [     -0.0662]
    [2]    'PDL13'    'JA2011'    [     -0.2187]
    [2]    'PDL13'    'SO2011'    [     -0.0745]

and I want to group all the observations that correspond to JF2008 (JF=January and February) as follows

 1  PDL12  JF2008   
2     'PDL13'    'JF2008'    [         NaN]

then all the observations that correspond to the next months

 [1]    'PDL12'    'MA2008'    [     -0.0330]
[2]    'PDL13'    'MA2008'    [      0.0218]

and so forth... . And then for the year 2009 similarly to 2008

thanks in advance

[Oleg Komarov]

Do you want to sort the table according to the third and then second row?

[antonet]

adrei bobrov gave the solution. thank you as well

Answer 1

[Andrei Bobrov]

eg

k = cellstr(reshape(datestr(datenum(0,1:12,1,0,0,0),'m'),2,[])')
t = cellfun(@(x){x(1:2) x(3:end)},A(:,3),'un',0)
tt = cat(1,t{:})
[a,a] = ismember(tt(:,1),k)
[id,id] = sortrows([a,str2double(tt(:,2))],[2,1])
out = A(id,:)

OR

in this is case

out = A(reshape(reshape(1:size(A,1),[],2)',[],1),:)

[antonet]

thank you. the first code fits my case

[antonet]

andrtei, I would also like to ask you something regarding the code

cellstr(reshape(datestr(datenum(0,1:12,1,0,0,0),'m'),2,[])')

you displayed above.

I want to do the same thing as before but for the following matrix

[1] 'PDL12' ' OCT 2008' [ 0.0073]

[1] 'PDL12' 'NOV 2008' [ 0.0100]

[1] 'PDL12' 'DEC 2008' [ 0.0149]

[1] 'PDL12' 'JAN 2009' [ 0.0081]

[1] 'PDL12' 'FEB 2009' [ 0.0055]

[1] 'PDL12' 'MAR 2009' [ 0.0099]

[1] 'PDL12' 'APR 2009' [ 0.0065]

[1] 'PDL12' 'MAY 2009' [ 0.0056]

[1] 'PDL12' 'JUN 2009' [ 0.0176]

[1] 'PDL12' 'JUL 2009' [ 0.0094]

[1] 'PDL12' 'AUG 2009' [ 0.0070]

[1] 'PDL12' 'SEP 2009' [ 0.0223]

[1] 'PDL12' 'OCT 2009' [ 0.0124]

[1] 'PDL12' 'NOV 2009' [ -0.0131]

[1] 'PDL12' 'DEC 2009' [ 0.0161]

[1] 'PDL12' 'JAN 2010' [ 0.0126]

[1] 'PDL12' 'FEB 2010' [ 0.0146]

[1] 'PDL12' 'MAR 2010' [ 0.0195]

[1] 'PDL12' 'APR 2010' [ 0.0378]

[1] 'PDL12' 'MAY 2010' [ 0.0414]

[1] 'PDL12' 'JUN 2010' [ 0.0356]

[1] 'PDL12' 'JUL 2010' [ 0.0276]

[1] 'PDL12' 'AUG 2010' [ 0.0860]

[1] 'PDL12' 'SEP 2010' [ 0.0258]

[1] 'PDL12' 'OCT 2010' [ 0.0354]

[1] 'PDL12' 'NOV 2010' [ 0.0375]

[1] 'PDL12' 'DEC 2010' [ 0.0366]

[1] 'PDL12' 'JAN 2011' [ 0.0203]

[1] 'PDL12' 'FEB 2011' [ -0.0078]

[1] 'PDL12' 'MAR 2011' [ -0.0158]

[1] 'PDL12' 'APR 2011' [ 0.0096]

[1] 'PDL12' 'MAY 2011' [ -0.0032]

[1] 'PDL12' 'JUN 2011' [ -0.0082]

[1] 'PDL12' 'JUL 2011' [ -0.0076]

[1] 'PDL12' 'AUG 2011' [ 0.0048]

[1] 'PDL12' 'SEP 2011' [ -0.0185]

[2] 'PDL13' ' OCT 2008' [ 0.0187]

[2] 'PDL13' 'NOV 2008' [ 0.0176]

[2] 'PDL13' 'DEC 2008' [ -0.0012]

[2] 'PDL13' 'JAN 2009' [ 0.0145]

[2] 'PDL13' 'FEB 2009' [ 0.0274]

[2] 'PDL13' 'MAR 2009' [ 0.0186]

[2] 'PDL13' 'APR 2009' [ -0.0106]

[2] 'PDL13' 'MAY 2009' [ -0.0034]

[2] 'PDL13' 'JUN 2009' [ 0.0230]

[2] 'PDL13' 'JUL 2009' [ 0.0231]

[2] 'PDL13' 'AUG 2009' [ 0.0030]

[2] 'PDL13' 'SEP 2009' [-7.0000e-004]

[2] 'PDL13' 'OCT 2009' [ 0.0015]

[2] 'PDL13' 'NOV 2009' [ -0.0241]

[2] 'PDL13' 'DEC 2009' [ 0.0227]

[2] 'PDL13' 'JAN 2010' [ 0.0133]

[2] 'PDL13' 'FEB 2010' [ 0.0110]

[2] 'PDL13' 'MAR 2010' [ 0.0142]

[2] 'PDL13' 'APR 2010' [ 0.0192]

[2] 'PDL13' 'MAY 2010' [ 0.0132]

[2] 'PDL13' 'JUN 2010' [ 0.0164]

[2] 'PDL13' 'JUL 2010' [ 0.0228]

[2] 'PDL13' 'AUG 2010' [ 0.0069]

[2] 'PDL13' 'SEP 2010' [ -0.0069]

[2] 'PDL13' 'OCT 2010' [ -0.0064]

[2] 'PDL13' 'NOV 2010' [ -0.0109]

[2] 'PDL13' 'DEC 2010' [ -0.0251]

[2] 'PDL13' 'JAN 2011' [ 0.0155]

[2] 'PDL13' 'FEB 2011' [ -0.0022]

[2] 'PDL13' 'MAR 2011' [ 0.0043]

[2] 'PDL13' 'APR 2011' [ 0.0132]

[2] 'PDL13' 'MAY 2011' [ 0.0113]

[2] 'PDL13' 'JUN 2011' [ -0.0026]

[2] 'PDL13' 'JUL 2011' [ 0.0011]

[2] 'PDL13' 'AUG 2011' [ 0.0131]

[2] 'PDL13' 'SEP 2011' [ 0.0142]

how can I do this using something similar to cellstr(reshape(datestr(datenum(0,1:12,1,0,0,0),'m'),2,[])')

thanks

[antonet]

actually i want to modify your code

k = cellstr(reshape(datestr(datenum(0,1:12,1,0,0,0),'m'),2,[])')

t = cellfun(@(x){x(1:2) x(3:end)},A(:,3),'un',0)

tt = cat(1,t{:})

[a,a] = ismember(tt(:,1),k)

[id,id] = sortrows([a,str2double(tt(:,2))],[2,1])

out = A(id,:)

but I have problem doing it

thanks

[Oleg Komarov]

Say the matrix you posted is called A:

[trash idx] = sortrows([cat(1,A{:,1}) datenum(A(:,3),'mmm yyyy')],[2 1]);

A(idx,:)

[antonet]

Thank you oleg. it works!