finding special date time

Hey everyone,

i working on the financial pattern recognition and i need to filter my extracted pattern by some date time method,

for instance, i need to preserve all patterns which occur on the same day in the entire dataset or preserve all pattern occur in the same month or in the same quarter of year.

(suppose that my dataset begins from 1-1-2017 till 1-1-2019 )

i think, if today is Sunday and i want to preserve all of Sunday in my dataset, the challenge is how to build a Date number Variable that contains all of Sundays from 1-1-2017 to 1-1-2019?! 

the second question is, in order to eliminate unuseful date from my DateTime Variable is there any work possible like logical indexing to speed up my code?

thanks in advanced

Abolfazl.

Answer 1

[Steven Lord]

Find the next Sunday after a given date.

T = datetime('today');
nextSun = dateshift(T, 'dayofweek', 'Sunday', 'next')

All Sundays between today (T) and the start of 2021.

start2021 = datetime(2021, 1, 1);
allSundays = nextSun:calweeks(1):start2021

Let's check.

[~, dayOfWeek] = weekday(allSundays)
% or
dName = day(allSundays, 'shortname')
dNum = day(allSundays, 'dayofweek')

Or if you already have a vector of datetime values, call day or weekday with one output (for day you'll want to specify 'dayofweek' as the kind) and extract those elements of your vector with day number 1 (Sunday.)

[Abolfazl Nejatian]

dear Steven,

thank you very much for your consideration.

i have find another way that insert it below,

nowDate     = datetime('today');                   % today Time
patternDate = datetime(2010,1,1):hours(.5):nowDate % let's suppose it is a time vector from 01/01/2000 with half hour interval

%---- finding the same month in our patternDate vector
Month = (patternDate.Month == nowDate.Month); 

%---- finding the same weekday in our patternDate vector
WeekDay = weekday(patternDate) == weekday(nowDate);


%---- pattern that happened in same month and same day of week
patternDate(WeekDay & Month)