# Should I use numpy and cython for complex numbers?

Anonymous

Here is the problem:

if I try this:

var('z')
z = 1+2*I
type(z)
<type 'sage.symbolic.expression.Expression'>


So, it isn't a complex number. BUT, if I do this:

I = CC.0
z = 1+2*I
type(z)


then I get a complex number

<type 'sage.rings.complex_number.ComplexNumber'>


Should I use numpy and cython for complex numbers?

edit retag close merge delete

Sort by » oldest newest most voted

It depends on what kind of complex number you want. Sage provide tons, you do not need numpy, cython or whatever. Here are some examples (these are not the only ones):

With

sage: z = 1+2*I
sage: z.parent()
Symbolic Ring


You indeed get symbolic representation of a complex number, so you can do symboloc things such as:

sage: exp(pi*z).simplify()
e^pi


WIth

sage: z = 1+2*CDF.0
sage: z.parent()
Complex Double Field


You get a fast numerical representation of your complex number (you should prefer CDF over CC since both have the same precision, but CDF takes the advantage of the CPU floating-point arithmetics, and the functions you will call from it will use optimized libraries).

With

sage: z = 1+2*ComplexIntervalField(1000).0
sage: z.parent()
Complex Interval Field with 1000 bits of precision


You get a certified numerical representation of your complex number with high precision (interval arithmetics).

With

sage: z = 1+2*QQbar.0
sage: z.parent()
Algebraic Field


You get an algebraic (hence exact) representation of your complex number (but pi does not exist here).

With

sage: z = 1+2*ZZ[i].1
sage: z.parent()
Gaussian Integers in Number Field in I with defining polynomial x^2 + 1


You get a Gaussian integer.

And so on...

Now, you just have to decide what do you want to do with z to select an appropriate representation.

more