469,299 Members | 2,106 Online

# Create linear spaced vector?

Hi Everyone,

I am trying to port some old MatLab code to python, and am stuck on
how to accomplish something.

I am trying to write a generalized function that will create a
linearly spaced vector, given the start and end point, and the number
of entries wanted.

In MatLab I have this function that I wrote:

Expand|Select|Wrap|Line Numbers
1.
2. function out = linearspace(x1,x2,n)
3.
4. out = [x1+ (0:n-2)*(x2 - x1)/(floor(n)-1) x2];
5.
6. return
7.
8.
9.

I have the numeric package, numarray installed, and I think it should
be accomplished easily, but I just can't seem to get the syntax
correct with python.

Any tips would be greatly appreciated.
Thanks
Jul 18 '05 #1
6 6261 On 2004-12-17, kjm <kj******@gmail.com> wrote:
I am trying to write a generalized function that will create a
linearly spaced vector, given the start and end point, and the number
of entries wanted.

from scipy import *
mgrid[0.0:10.0:5j]

array([ 0. , 2.5, 5. , 7.5, 10. ])

--
Grant Edwards grante Yow! Is this where people
at are HOT and NICE and they
Jul 18 '05 #2
On Fri, 2004-12-17 at 13:39, kjm wrote:
Hi Everyone,

I am trying to port some old MatLab code to python, and am stuck on
how to accomplish something.

I am trying to write a generalized function that will create a
linearly spaced vector, given the start and end point, and the number
of entries wanted.

In MatLab I have this function that I wrote:

Expand|Select|Wrap|Line Numbers
1.  function out = linearspace(x1,x2,n)
2.  out = [x1+ (0:n-2)*(x2 - x1)/(floor(n)-1) x2];
3.  return
4.

I have the numeric package, numarray installed, and I think it should
be accomplished easily, but I just can't seem to get the syntax
correct with python.

Any tips would be greatly appreciated.
Thanks

Is this want you want?

#!/usr/bin/python

def linear_space( start, end, count ):

""" Returns a vector containing count evently spaced intervals
(count + 1 evenly spaced points) """

delta = (end-start) / float(count)
return [start,] + \
map( lambda x:delta*x + start, range( 1, count ) ) + [end, ]

if __name__ == "__main__":
print linear_space( 1.0, 2.0, 10 )

Running it gives you:
[1.0, 1.1000000000000001, 1.2, 1.3, 1.3999999999999999, 1.5,
1.6000000000000001, 1.7000000000000002, 1.8, 1.8999999999999999, 2.0]

Jul 18 '05 #3
>>>>> "kjm" == kjm <kj******@gmail.com> writes:

kjm> Hi Everyone, I am trying to port some old MatLab code to
kjm> python, and am stuck on how to accomplish something.

kjm> I am trying to write a generalized function that will create
kjm> a linearly spaced vector, given the start and end point, and
kjm> the number of entries wanted.

kjm> In MatLab I have this function that I wrote:

kjm> [code]

kjm> function out = linearspace(x1,x2,n)

in matlab the builtin function to accomplish this is "linspace"

The python package matplotlib defines a host of matlab compatible
functions, including linspace

def linspace(xmin, xmax, N):
if N==1: return xmax
dx = (xmax-xmin)/(N-1)
return xmin + dx*arange(N)
Note that matplotlib extends the Numeric/numarray core of matlab
compatible functions (defined in MLab) to include plotting functions

http://matplotlib.sourceforge.net

A listing of matlab compatible functions is provided at
http://matplotlib.sourceforge.net/matplotlib.pylab.html
JDH

Jul 18 '05 #4
Thanks for the code snippets guys. Exactly what I needed to get going.

I knew I could get the solution from matplotlib, but getting it
installed using Fink (OS X) has been giving me a headache, so I thought
I could just write my own function for now to get a small piece of code
written....

The help is greatly appreciated.

kjm

John Hunter wrote:
>> "kjm" == kjm <kj******@gmail.com> writes:

kjm> Hi Everyone, I am trying to port some old MatLab code to
kjm> python, and am stuck on how to accomplish something.

kjm> I am trying to write a generalized function that will create
kjm> a linearly spaced vector, given the start and end point, and
kjm> the number of entries wanted.

kjm> In MatLab I have this function that I wrote:

kjm> [code]

kjm> function out = linearspace(x1,x2,n)

in matlab the builtin function to accomplish this is "linspace"

The python package matplotlib defines a host of matlab compatible
functions, including linspace

def linspace(xmin, xmax, N):
if N==1: return xmax
dx = (xmax-xmin)/(N-1)
return xmin + dx*arange(N)
Note that matplotlib extends the Numeric/numarray core of matlab
compatible functions (defined in MLab) to include plotting functions

http://matplotlib.sourceforge.net

A listing of matlab compatible functions is provided at
http://matplotlib.sourceforge.net/matplotlib.pylab.html
JDH

Jul 18 '05 #5
>>>>> "kjmacken" == kjmacken <kj******@gmail.com> writes:

kjmacken> Thanks for the code snippets guys. Exactly what I
kjmacken> needed to get going. I knew I could get the solution
kjmacken> from matplotlib, but getting it installed using Fink (OS
kjmacken> X) has been giving me a headache, so I thought I could
kjmacken> just write my own function for now to get a small piece
kjmacken> of code written....

Yes, matplotlib fink installs have frustrated many an OSX user. Note
that the matplotlib.mlab package (where linspace and others functions
reside) do not require any extension code and can be reused anywhere
you like as python code by copying and pasting, etc.

Also, Robert Kern is in the process of building an "enthon" package
for OSX that has most of the utilities for scientific computing
including matplotlib built-in. Batteries included on steroids,
basically.

http://www.scipy.org/wikis/featurerequests/MacEnthon

So keep your eyes on that site for release information.

JDH
Jul 18 '05 #6
John,

Thanks for the heads up RE: scipy, I will keep my eyes on the
developments.

Also, thanks for the info:

that the matplotlib.mlab package (where linspace and others functions
reside) do not require any extension code and can be reused anywhere
I was able to get these modules into my site-packages directory, and
make use of what is there.

Cheers,

kjm

Jul 18 '05 #7

### This discussion thread is closed

Replies have been disabled for this discussion.