matlab code

[Willy26]

Hi and thank you for considering this post.
I have 2 time series (MID1 and MID2). They have a table format with 2 rows each. The first row is the date and the second row is a value. Both time series don't have continuouse dates.

For example
MID1:
01.02.2010 - 34
04.02.2010 - 37
07.02.2010 - 60
08.02.2010 - 61

MID2:
31.01.2010 - 20
02.02.2010 - 30
04.02.2010 - 32
06.02.2010 - 41
07.02.2010 - 40

I want to compare the data row. If the time series of MID1 and MID2 have the same date (in the example the 04.02.2010 and 07.02.2010) the values in the value rows should be summarized. If the comparison don't find the same date, it should be considered just the value of MID1 or MID2.
in the case of the example like this:

total:
31.01.2010 - 20
01.02.2010 - 34
02.02.2010 - 30
04.02.2010 - 37 + 32 = 69
06.02.2010 - 41
07.02.2010 - 60 + 40 = 100
08.02.2010 - 61

I was thinking it could work with a double loop like this:

MID1 with 1st row (date) and 2nd row (value)
MID2 with 1st row (date) and 2nd row (value)

-------------
total=zeros(2709,1)
for i=1:2709
for ii=1:2709
if MID1.date(i)==MID2.date(ii)
total(i)= MID1.value+MID2.value
elseif MID1.date(i)~=MID2.date(ii)
total(i)=MID1.value
else MID2.date(i)~=MID1.date(ii)
total(i)=MID2.value

end
end
end
-------------

As I'm super weak in programming, it didn't work out well. Could you help me in solving the problem?

thanks a lot!
​

 

[GregLocock]

I'm afraid your logic is wrong. There is no guarantee that MID1 has the same number of values as MID2. You are assuming that MID1 has all of MID2's dates in it. I suggest you start with a small subset of your data, and print the results out.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?