How to insert zeros in a data returns

I have a data with two variables: dates and returns. These two variables don't include the festivities, i'd like to insert zeros in the returns in those days. Is there anyone who can help me?

12 个评论

I tried to use datenum to have serial consequent number that are missing when there are festivities but i can't continue at the moment.
Could you provide an example of what you'd like to do?
Dates(3x1)=[2013-11-01, 2013-11-02, ( jump saturday and sunday), 2013-11-05]
And these are the related returns for these days:
Returns(3x1)=[1.05, 1,03, (jump), 1.05]
Finalreturnmatrix(5x1)=[1.05, 1.03, 0, 0, 1.05]
Consider that my data jumps also other festivities (like christmas etc..).
I need only a vector like "finalreturnmatrix", the vector of the dates is not required to be completed. I hope it's more understandable now. Thank you
Yes, it's more understandable.
What format are you "Dates"? Are they datetime? Are they strings?
class(Dates) = ?
"cells" or "double" if i import them as a date/time. At the moment it says :
"Undefined function 'datetime' for input arguments of type 'double'.
Error in changeforma (line 2)
t = datetime(Dates,'InputFormat','yyyy-MM-dd');"
Thanks for your patience.
If you import them as datetime, you can name that variable 't' (for now) and skip this line of code from my answer:
t = datetime(Dates,'InputFormat','yyyy-MM-dd');
If you continue to have problems, please attach a sample of data instead of me guessing what you're working with.
I've got the r2014a version of matlab, it seems that i can't use the functions that you wrote before. There aren't in my documentation. Anyway i uploaded a datasample, there are prices instead returns but it's the same thing.
The csv file is corrupted (when I opened it it was gibberish). Anyway, see my updated answer; at the bottom is a low-level solution that doesn't involve any newer matlab features.
When i use your code as in your example it works but, when i use it with all of my data, it says:
"In an assignment A(I) = B, the number of elements in B and I must be the same.
allReturns(ismember(alldates,Dates)) = Returns;"
Consider that the variables have the same dimensions. I also uploaded another sample in a different format if it's useful
I just updated my answer to fix a mistake in my date formats. I used capital MM (minutes) instead of lowercase mm (months). Fixed (I tested it with your data).
It works! I really appreciated your help, good job.

请先登录,再进行评论。

 采纳的回答

For r2016b or later
Assuming your dates are a cell array of strings in the format you described in the comments under your question, you can put your data into a timetable, then use retime() to fill in the missing dates.
Dates = {'2013-11-01', '2013-11-02', '2013-11-05', '2013-11-06'};
Returns = [1.05, 1.03, 1.05, 1.03];
% Convert dates to datetime
t = datetime(Dates,'InputFormat','yyyy-MM-dd');
% put dates and data into timetable
tt = timetable(t',Returns','VariableName',{'Returns'});
% expand your time table to include all missing days; fill missing Returns with 0
ttFill = retime(tt,'daily','fillwithconstant','Constant',0);
% Extract the vector (if you must)
Finalreturnmatrix = ttFill.Returns;
Results
ttFill =
6×1 timetable
Time Returns
___________ _______
01-Nov-2013 1.05
02-Nov-2013 1.03
03-Nov-2013 0
04-Nov-2013 0
05-Nov-2013 1.05
06-Nov-2013 1.03
>> Finalreturnmatrix
Finalreturnmatrix =
1.05
1.03
0
0
1.05
1.03
Lower level version (for earlier releases)
'allReturns' is the vector you're looking for. This approach doesn't rely on datetime, timetables, etc.
Dates = {'2013-11-01', '2013-11-02', '2013-11-05', '2013-11-06'};
Returns = [1.05, 1.03, 1.05, 1.03];
alldates = datestr(datenum(Dates{1},'yyyy-mm-dd') : 1 : datenum(Dates{end},'yyyy-mm-dd'),'yyyy-mm-dd');
allReturns = zeros(size(alldates,1),1);
allReturns(ismember(alldates,Dates)) = Returns;

更多回答(0 个)

类别

帮助中心File Exchange 中查找有关 Time Series Objects 的更多信息

产品

版本

R2014b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by