Save real numbers (with decimals) in a loop into vector.

Dear All, I asked a question yesterday, but I wasn't so clear and believe that the information in the post got complicated and many would avoid looking at it~ I rephrase my problem very shortly:

Normally, if 'i' is a vector of positive integers, the vector of its product say 'i*rand' would be easy to get:

      i=1:5;
      y(i)=i*rand

y =

        0.4468    0.8935    1.3403    1.7871    2.2339

But if 'i' is just a real number, say:

      i=0.1:0.5;
      y(i)=i*rand

Attempted to access y(0.1); index must be a positive integer or logical.

Question:

How do I overcome this problem and get 'y', the vector of the product 'i*rand'.

Thanks and Best of Regards// Ali

Answer 1

[Azzi Abdelmalek]

ii=0.1:0.5;
counter=1:numel(ii)
y(counter)=ii*rand

[AND]

Dear Azzi,

Many Many Thanks for the support//

I think I am almost there, but when I use the answer you just provided - I get just a scalar, that is just one value for y which stands for ii=0.1

    ii = 

        0.1000

    y =

        0.0268

Any suggestions?

[Azzi Abdelmalek]

Sorry, it's

ii=0.1:0.1:0.5;
counter=1:numel(ii)
y(counter)=ii*rand

[AND]

Dear Azzi, That did the trick for this, and so that is the answer - but it doesn't work with my function, I tried

s=str2double(get(handles.StockP,'String'));
 counter=1:numel(s);
y(counter)=s

but it is giving again:

y =

       101

    y =

      114.9920

    y =

      116.1000

    y =

      113.3770

    y =

       98.3933

If you help me overcome this, you would have saved my day :)

Thanks again//

[Azzi Abdelmalek]

It seems that the length of s is 1. What are you then expecting to have? Can you explain what is s and its size?

[AND]

s is a stock price.

Suppose I set the number of simulation to 5:

At the beginning of the simulation, the first function will see if the number is > 0, set new value of number = n-1 (5-1) and call the second function which generates new stock price. At the end of the stock generating function, I call the first function and so the process will loop until n=0. Thus If I set n to 5, I would have 5 loops and so 5 different stock prices.

What I need is to record these prices in order to get averages, plot them and so on.

I tried several methods, some of which you proposed but they didn't work so far. The main idea is to get the last stock price in every loop, I do this by retrieving the string

s=str2double(get(handles.StockP,'String'))

but what I am not able to do is to put all these different prices in one vector instead of having them as separate entities such as

s =

    117.0050

s =

    100.0460

s =

     96.5053

s =

     99.0834

s =

    114.7970

I need the result to be:

s =

    117.0050  100.0460  96.5053  99.0834  114.7970

Any suggestions?

[Azzi Abdelmalek]

s=str2double(get(handles.StockP,'String'))
y=get(handles.StockP,'userdata')
y=[y s]
set(handles.StockP,'userdata',y)

[AND]

Dear Azzi, Thanks heaps for your generous help// God Bless You

I can say that you saved my day :) and so the other question I raised yesterday is solved too.

The only thing is that it saves the array and everytime I start the function it adds the new values to the old values, but managed to overcome this by reseting the GUI.

Thanks again and all best// Ali

Answer 2

[Jan]

You can simply omit the index on the left hand side:

ii = 0.1:0.5;
y  = ii * rand

After this, you cannot access "y(0.1)", because index expressions must be integers. But you could create a function:

function r = y(x)
r = rand * x;

Now "y(0.1)" calculates the function for this value, but this is not exactly equivalent to the indexing in your example. But usually this is a sufficient solution.

When you really need to use a floating point value as index, you will get problems with FAQ: limited floating point precision:

any(0:0.1:1 == 0.3)
>> 0

This is not a bug, but cause by the fact, that not all decimal numbers can be represented exactly in binary format. Therefore any kind of floating-point-indexing would require a comparison like find(abs(v - value) < 100*eps). The limit of 100*eps is more or less arbitrary and depend on the values. In consequence the required code for indexing must be smart and need intelligent fallbacks.

Better stay on integer indexing.

[Seban]

You can then simply get an index vector with the same length as the non-integer indeces:

str = {'.1' '.2' '.3' '.4' '.5'};
ii  = str2double(str);
y   = ii*rand
ind = 1:length(y)

[AND]

Thanks for your contribution Jan, much appreciated.

The function method does make sense, but in my particular case --> simulating stock price (please refer to my last comment), I tried creating the

function r=y(s)

where s, stock price, is the variable I am looking for, then I call the function y(s) in every loop, but the result is the same, just a scalar not a vector:

s =

117.0050

s =

100.0460

s =

   96.5053

s =

   99.0834

s =

114.7970

I wonder if you have any suggestions of how to put all these values into one vector?