Question

I am trying to grab the max for each year in my dataset. The data is in one column, in order at a daily timestep.
How do I get the max from the first 365 rows, then from the next 365 rows?
I was thinking something like:

years=31;
for i=1:years
peak(i)=max(data(i:365*i,2))
end

but when i=2, the range should be from 366:730 and so forth.
The data matrix is 11322×7 double and I need column 2.

Answer 1

I thought about suggesting something like @Thilo’s answer, but that doesn’t take into account leap years (which appear to show up in your data, since 365*31=11315, which is less than 11322)

You might be able to gin up a complicated vectorized solution, but I’d be tempted to just manually keep track of the start and stop indices:

NYEARS = 31;

start = 1;
stop = 365 + isLeapYear(year(1));
n=1;
maxValues = nan(NYEARS,1);
while(n=<NYEARS)
  maxValues(n) = max(data(start:stop,2));
  n=n+1;
  start = stop + 1;
  stop = start + 365 + isLeapYear(year(n));
end

It’s probably not blazingly fast, but I doubt this is going to be a bottleneck either.

function leap_p = isLeapYear(year)
  leap_p = ~mod(year,400) || (~mod(year,4) && mod(year,100)) 
end