Ignore NaN values in a Table

Hello - apologes if this has been covered previously, but i can't find a satisfactory answer; any help would be greatly appreciated!

I have a table, called "returns" (% performance for some investments), see below. Some cells are NaN.

I wish to compute a range of statistics based on these numbers: annualised return / annualised volatiity / contribution to return / contribution to volatility / contribution to drawdown etc.

Date return1 return2 return3 return4

....

'30-Apr-2020' 0.056 0.0097 0.0467 0.0240

'31-May-2020' NaN 0.0152 0.0635 0.0450

How can i ignore this NaN value? For example, if i take the example of contribution to return, with the code below. Simply put, how do I ignore the NaN when calculateing the annualized return for each (the first section, highlighed in bold). This code below will return annualized return for funds 2, 3 and 4, but for fund 1, with the NaN - it will not work.

%% CONTRIBUTION TO RETURN

% annualized returns for each fund I

I = (prod(returns+1).^(12/length(returns)))-1;

% weighted annualized return

I_W = I .* weights;

% portfolio returns, monthly and annualized

% amended: portfolio returns taken direct from SP - appropriate rebalancing

% of portfolio when fund track not available

%monthly_fund_contrib = returns .* weights;

%portfolio_return = sum(monthly_fund_contrib')';

portfolio_return = outputdata2(:,2); %ensure this is double

portfolio_return = cell2mat(portfolio_return);

R = (prod(portfolio_return+1).^(12/length(portfolio_return)))-1;

% difference between portfolio return and sum of contributions

Diff = R - sum(I_W);

% weight this difference across funds in the portfolio

Diff_W = weights * Diff;

Return_Contribution = I_W + Diff_W;

Return_Contribution = array2table(Return_Contribution);

many thanks for any help!

Answer 1

[Star Strider]

Since the Question appears to apply to the prod funciion, see the section on Product Excluding NaN. (I am not certain when that option was introduced, however it is present in all recent versions/releases.)

.

Answer 2

[Lalitha Tejaswi]

The functions prod and sum have an option to provide the nanflag where you can include the NaNs (use 'includenan') or omit the nans ('omitnan') while you use these functions.

Example: For computing the product of the following array elements without considering NaNs,

              x = [1 2 3 NaN];
              product = prod(x,'omitnan')

product = 6

Same goes with sum. Use this option for your computations.

Hope this helps!

[Nicholas Thomason]

fantastic - many thanks for your help!