'Typecast to an int in Octave/Matlab

I need to call the index of a matrix made using the linspace command, and based on some data taken from an oscilloscope. Because of this, the data inputed is a double. However, I can't really call:

Time[V0Found]

where V0Found is something like 5.2 however, taking index 5 is close enough, so I need to drop the decimal. I used this equation to drop the decimal:

V0FoundDec = V0Found - mod(V0Found,1)
Time[V0FoundDec]

However, even though that drops the decimal, octave still complains about it.

So, what can I do to typecast it to an int?

arraysmatlabindexingoctavecasting


Solution 1:[1]

In MATLAB, it should be int8(x) or int16(x) or one of the other integer casts.

But I'm surprised you need to do that for an index. Try 

myarray(floor(indexlist))

or

myarray(round(indexlist))

where myarray is your array and indexlist is your vector of noninteger indices.

example:

octave-3.2.3:8> v=rand(1,8)*10+1
v =

   3.1769   1.4397   8.7504   1.7424   6.9413   3.1663   8.4085   9.0179

octave-3.2.3:9> a = (1:1:20).^2
a =

 Columns 1 through 15:

     1     4     9    16    25    36    49    64    81   100   121   144   169   196   225

 Columns 16 through 20:

   256   289   324   361   400

octave-3.2.3:10> a(floor(v))
ans =

    9    1   64    1   36    9   64   81

Solution 2:[2]

You could use round, floor, ceil functions instead of your formula to do the rounding.

By the way, indexing is done using parenthesis instead of brackets so:

V0FoundDec = round(V0Found);
Time(V0FoundDec) % not Time[V0FoundDec]

In case that was your problem

Solution 3:[3]

In Matlab, the right way to do this is to use the interp1 function to interpolate. The format of this command is

yout = interp1 (xdata, ydata, xin, ...)

or

yout = interp1 (ydata, xin, ...)

where xdata is then assumed to be 1:length(ydata).

If you want to produce the result you would get from calling

V0FoundDec = Time(round(V0found))

you would say

V0FoundDec = interp1(Time, V0found, 'nearest')

but you can just as easily get linear interpolation (this is the default) with

V0FoundDec = interp1(Time, V0found)

or

V0FoundDec = interp1(Time,V0found,'linear')

and you can also extrapolate outside the limits (using 'extrap' or providing an extrap value), where

Time(round(V0found))

will crash if round(V0found) < 1 or > length(Time)

Solution 4:[4]

For those using Octave, the cast function does the job:

octave:1> cast(5.2,'int8')
ans = 5

Sources

This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.

Source: Stack Overflow

Solution Source
Solution 1
Solution 2 Azim J
Solution 3 Cris Luengo
Solution 4 alek

Related Questions