how to calculate common dates from two date arrays?

I have two arrays with many datenums in format [startTime endTime]. I am trying to find a way to calculate all common datetimes (i.e duration), and end product should be new array where is all [startTime endTime] of all common datetimes.

Example (datenums are datetime for clarity in this example): dates1 array some row is :

[1.10.2025 10:00:00, 1.10.2025 13:00:00] 

and some rows in dates2 are:

[1.10.2025 09:30:00, 1.10.2025 11:00:00] 
[1.10.2025 12:30:00 ,1.10.2025 14:40:00] 

so new array of common times will get rows

[1.10.2025 10:00:00, 1.10.2025 11:00:00] 
[1.10.2025 12:30:00, 1.10.2025 13:00:00] 

Tried to ask Grok but its function was erroneus. Greatly appreciate help!

[Dyuman Joshi]

load('dates.mat')
whos

  Name          Size            Bytes  Class     Attributes

  ans           1x33               66  char                
  dates1      339x1             86825  struct              
  dates2      351x1             89897  struct              

dates1

dates1 = 339×1 struct array with fields:
    dStrt
    dStp

y = dates1.dStrt;
class(y)

ans = 'double'

y(1)

ans = 7.3933e+05

unique(y)

ans = 7.3933e+05

z= dates1.dStp;
class(z)

ans = 'double'

z(1)

ans = 7.3933e+05

unique(z)

ans = 7.3933e+05

Note - The dates1 data you have attached contains numeric double values (with all the values being same), not datenums - as you can see above.

I'm not sure how to interpret this.

[Sven Larsen]

interesting. they are all datenums and for example unique([dates1.dStrt]); gives 339 unique values....

but it is not relevant are they datenums or numbers. same algorithm works finding common number ranges also :)

[Walter Roberson]

datenum(now)

ans = 7.3991e+05

This is tthe same range as the 7.3933e+05 shown above, so the values are indeed datenum. The values are not datetime objects.

[Dyuman Joshi]

"they are all datenums and for example unique([dates1.dStrt]); gives 339 unique values...."

Yes, it escaped me that they can be datenums().

It seems I am missing something -

load('dates.mat')
whos

  Name          Size            Bytes  Class     Attributes

  ans           1x33               66  char                
  dates1      339x1             86825  struct              
  dates2      351x1             89897  struct              

dates1

dates1 = 339×1 struct array with fields:
    dStrt
    dStp

y = dates1.dStrt;
size(y)

ans = 1×2

1 1

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

z = vertcat(dates1.dStrt);
size(z)

ans = 1×2

339 1

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

"same algorithm works finding common number ranges also :)"

Similar logic will work. Same algorithm might not.

Please check Star Strider's answer below.

[dpb]

@Sven Larsen - convert the datenum arrays to datetimes and then see isbetween

Answer 1

[dpb]

load dates
clear ans
dates1=[datetime(vertcat(dates1.dStrt),'convertfrom','datenum') datetime(vertcat(dates1.dStp),'convertfrom','datenum')];
tD1=array2table(dates1,'VariableNames',{'Start','Stop'});
tD1.Length=tD1.Stop-tD1.Start;
head(tD1)

           Start                    Stop             Length  
    ____________________    ____________________    _________

    19-Mar-2024 16:10:53    19-Mar-2024 16:10:45    -00:00:08
    21-Mar-2024 06:06:46    19-Mar-2024 16:10:53    -37:55:53
    21-Mar-2024 08:28:59    21-Mar-2024 06:06:46    -02:22:13
    23-Mar-2024 22:25:16    21-Mar-2024 08:28:59    -61:56:17
    24-Mar-2024 11:33:38    23-Mar-2024 22:25:16    -13:08:22
    24-Mar-2024 16:57:32    24-Mar-2024 11:33:38    -05:23:54
    24-Mar-2024 16:59:15    24-Mar-2024 16:57:32    -00:01:43
    24-Mar-2024 20:42:44    24-Mar-2024 16:59:15    -03:43:29

Ooops!!! The start/stop times appear to be reversed as the length of each time span is negative.,,,we'll just recreate the table as expected.

tD1=array2table(dates1,'VariableNames',{'Stop','Start'});
tD1.Length=tD1.Stop-tD1.Start;
head(tD1)

            Stop                   Start             Length 
    ____________________    ____________________    ________

    19-Mar-2024 16:10:53    19-Mar-2024 16:10:45    00:00:08
    21-Mar-2024 06:06:46    19-Mar-2024 16:10:53    37:55:53
    21-Mar-2024 08:28:59    21-Mar-2024 06:06:46    02:22:13
    23-Mar-2024 22:25:16    21-Mar-2024 08:28:59    61:56:17
    24-Mar-2024 11:33:38    23-Mar-2024 22:25:16    13:08:22
    24-Mar-2024 16:57:32    24-Mar-2024 11:33:38    05:23:54
    24-Mar-2024 16:59:15    24-Mar-2024 16:57:32    00:01:43
    24-Mar-2024 20:42:44    24-Mar-2024 16:59:15    03:43:29

dates2=[datetime(vertcat(dates2.dStrt),'convertfrom','datenum') datetime(vertcat(dates2.dStp),'convertfrom','datenum')];
tD2=array2table(dates2,'VariableNames',{'Stop','Start'});
tD2.Length=tD2.Stop-tD2.Start;
head(tD2)

            Stop                   Start             Length 
    ____________________    ____________________    ________

    19-Mar-2024 16:10:53    19-Mar-2024 16:10:43    00:00:10
    21-Mar-2024 06:06:32    19-Mar-2024 16:10:53    37:55:39
    21-Mar-2024 08:29:13    21-Mar-2024 06:06:32    02:22:41
    23-Mar-2024 22:25:14    21-Mar-2024 08:29:13    61:56:01
    24-Mar-2024 00:33:46    23-Mar-2024 22:25:14    02:08:32
    24-Mar-2024 02:00:09    24-Mar-2024 00:33:46    01:26:23
    24-Mar-2024 11:33:29    24-Mar-2024 02:00:09    09:33:20
    24-Mar-2024 20:42:46    24-Mar-2024 11:33:29    09:09:17

Now it's not quite clear to me what you actually want as a result -- for each start time in tD1 the records that are in the duration of that time span whose start time is within that length of time or which overlaps or ...?

Can finish up once have a clear definition.

ixoverlap=arrayfun(@(t_s,t_e)(t_s<=tD2.Stop)&(t_e>=tD2.Start),tD1.Start,tD1.Stop,'uni',0);
ix=ixoverlap{1};
for i=2:numel(ixoverlap)
  ix=ix|ixoverlap{i};
end
res=tD1(ix,:);
height(res)

ans = 10

res

res = 10×3 table
            Stop                   Start             Length 
    ____________________    ____________________    ________

    19-Mar-2024 16:10:53    19-Mar-2024 16:10:45    00:00:08
    21-Mar-2024 06:06:46    19-Mar-2024 16:10:53    37:55:53
    21-Mar-2024 08:28:59    21-Mar-2024 06:06:46    02:22:13
    23-Mar-2024 22:25:16    21-Mar-2024 08:28:59    61:56:17
    24-Mar-2024 11:33:38    23-Mar-2024 22:25:16    13:08:22
    24-Mar-2024 16:57:32    24-Mar-2024 11:33:38    05:23:54
    24-Mar-2024 16:59:15    24-Mar-2024 16:57:32    00:01:43
    24-Mar-2024 20:42:44    24-Mar-2024 16:59:15    03:43:29
    25-Mar-2024 09:16:45    24-Mar-2024 20:42:44    12:34:01
    27-Mar-2024 16:26:51    25-Mar-2024 09:16:45    55:10:06

I THINK the above will find those that are overlapping; check the results to see if are what you were looking for.

[dpb]

ERRATUM/ADDENDUM

NOTA BENE: I forgot to remove the 1:10 debugging short subset subscripts on the arguments in the arrayfun() line above; I just made the edit to remove but the result will be many more than 10, I'm certain.