Extract rows from a matrix using time ranges

Hello everyone, i have a problem. I have to extract rows from a matrix using time ranges (each row between those dates), for example

mi matrix look like,

  A=734139    2         1.2 
      734140    3          1.4
      734141    10        0.4
      734142    8          0.1
      734143    14        2.1
      734144    20        4.3   
      734145    1          5.1
      734146    2          3.1
      734147    3          2.1
      734148    2          1.3

And i have other matrix with time ranges for example (initial date and final date of ranges):

B= 734139 734141
     734145  734145

I expect have this

C= 734139    2         1.2 
     734140    3          1.4
      734141    10        0.4
     734145    1          5.1
      734146    2          3.1

I know how did it with the new time tools (like timetables for example).. in fact is done, but i must use only basic tools because i'm translatting this code for old products (matlab2012a specifically) and octave.

Thank you for your help!

Answer 1

[Rik]

I'm going to assume your last element of B should be 1 larger, because then your C makes sense. The code below should do what you need by looping over your different time segments.

A= [734139   2   1.2
    734140    3   1.4
    734141   10   0.4
    734142    8   0.1
    734143   14   2.1
    734144   20   4.3
    734145    1   5.1
    734146    2   3.1
    734147    3   2.1
    734148    2   1.3];
B= [734139 734141
    734145 734146];

row=false(size(A,1),1);
t_=A(:,1);
for k=1:size(B,1)
    t1=B(k,1);t2=B(k,2);
    row=row | (t_>=t1 & t_<=t2);
end
C=A(row,:);
clc
fprintf('%6d  %02d  %3.1f\n',C')

[Adam Danz]

Similar solution that stores each comparison in a matrix and then combines columns.

d = false(size(A,1),size(B,1)); 
for i = 1:size(B,1)
    d(:,i) = A(:,1)>=B(i,1) & A(:,1)<=B(i,2); 
end
finalSelection = A(any(d,2),:); 

[Rik]

Thanks for the inspiration Adam. Because of the structure of your code I had a look at implicit expansion again. The code below is your loop, but without a loop:

d=bsxfun(@ge,A(:,1),B(:,1)') & bsxfun(@le,A(:,1),B(:,2)');
C=A(any(d,2),:);
clc,fprintf('%6d  %02d  %3.1f\n',C')

[Adam Danz]

Nice teamwork! :D

[duilio fonseca]

Thank you both! i'll try

Answer 2

[Peter Perkins]

dulio, I strongly recommend that you use datetims and timetables. You will likely be happier in the long run:

>> X = [734139   2   1.2
    734140    3   1.4
    734141   10   0.4
    734142    8   0.1
    734143   14   2.1
    734144   20   4.3
    734145    1   5.1
    734146    2   3.1
    734147    3   2.1
    734148    2   1.3];
>> t = array2table(X,'VariableNames',{'Date' 'X' 'Y'});
>> t.Date = datetime(t.Date,'ConvertFrom','datenum','Format','dd-MMM-yyyy');
>> tt = table2timetable(t)
tt =
  10×2 timetable
       Date        X      Y 
    ___________    __    ___
    01-Jan-2010     2    1.2
    02-Jan-2010     3    1.4
    03-Jan-2010    10    0.4
    04-Jan-2010     8    0.1
    05-Jan-2010    14    2.1
    06-Jan-2010    20    4.3
    07-Jan-2010     1    5.1
    08-Jan-2010     2    3.1
    09-Jan-2010     3    2.1
    10-Jan-2010     2    1.3
>> dt = datetime([734139 734141],'ConvertFrom','datenum')
dt = 
  1×2 datetime array
   01-Jan-2010 00:00:00   03-Jan-2010 00:00:00
>> tr = timerange(dt(1),dt(2))
tr = 
	timetable timerange subscript:
		Select timetable rows with times in the half-open interval:
		[01-Jan-2010 00:00:00, 03-Jan-2010 00:00:00)
	See Select Timetable Data by Row Time and Variable Type.
>> tt1 = tt(tr,:)
tt1 =
  2×2 timetable
       Date        X     Y 
    ___________    _    ___
    01-Jan-2010    2    1.2
    02-Jan-2010    3    1.4

[Adam Danz]

datetime() was released in r2014b. table2timetable() and timerange() in r2016b. OP mentioned that this must work in r2012a.

[duilio fonseca]

Hi Peter, thank you very much in fact i'm using those functions in my original code. But how Adams writes, i have to run this in r2012a product.

[Peter Perkins]

Got it, I did not catch that. Sorry.