How to write a data file in Python

javiergra24

Hi guys

I've done a Runge-Kutta script for the Lorenz Equations in python, I need to write data for (t,x,y,z) in a .dat file in order to reprint in a table. How can a do it?

Imported routines are:

import sys, pylab, numpy
from pylab import *
from mpl_toolkits.mplot3d import Axes3D
from numpy import *
import matplotlib.pyplot as plt

Mi iteration code is:

for n in xrange(0,N):
# at each time step calculate new x(t),y(t),z(t)
# and append to lists x0, y0, z0
x,y,z = RKTwoD(x0[n],y0[n],z0[n],VDPXDot,VDPYDot,VDPZDot,dt)
x0.append(x)
y0.append(y)
z0.append(z)
t.append(t[n] + dt)
savetxt('output.dat', (t,x0,y0,z0))

But it doesn't work, file just contain six columns instead of four and it takes too long to execute the script.

Please can anyone help me?

Thank you

story645

If you keep everything as an array instead of a list, basically take
x0.append(x)
y0.append(y)
z0.append(z)

and do a vstack or hstack at the end
so table = hstack(x0,y0,z0)
then you just need to write out
table.tofile('output.dat')

numpy lists don't write out straight to dats

javiergra24

Thanks for your answer but it doesn't work

I'm trying

Code:

import sys, pylab, numpy
from pylab import *
from mpl_toolkits.mplot3d import Axes3D
from numpy import *
import matplotlib.pyplot as plt
#I'll use values of sigma = 6.0, beta = 3.0 and rho = 36
# The Lorenz Equations are dx/dt=sigma*(y-x) ; dy/dt=x*(rho-z)-y ; dz/dt=x*y-beta*z
def VDPXDot(x,y,z):
	return 6.0 * (y-x)

def VDPYDot(x,y,z):
	return x * (40.0 - z)- y

def VDPZDot(x,y,z):
	return x * y - 3.0 * z

# 3D Fourth-Order Runge-Kutta Integrator
def RKTwoD(x,y,z,f,g,m,dt):
	#9 operaciones
        k1x = dt * f(x,y,z)
	k1y = dt * g(x,y,z)
        k1z = dt * m(x,y,z)
	#15 operaciones
        k2x = dt * f(x + k1x / 2.0,y + k1y / 2.0,z + k1z / 2.0)
	k2y = dt * g(x + k1x / 2.0,y + k1y / 2.0,z + k1z / 2.0)
        k2z = dt * m(x + k1x / 2.0,y + k1y / 2.0,z + k1z / 2.0)
        #15 operaciones
        k3x = dt * f(x + k2x / 2.0,y + k2y / 2.0,z + k2z / 2.0)
	k3y = dt * g(x + k2x / 2.0,y + k2y / 2.0,z + k2z / 2.0)
        k3z = dt * m(x + k2x / 2.0,y + k2y / 2.0,z + k2z / 2.0)
	#12 operaciones
        k4x = dt * f(x + k3x,y + k3y,z + k3z)
	k4y = dt * g(x + k3x,y + k3y,z + k3z)
        k4z = dt * m(x + k3x,y + k3y,z + k3z)
        #12 operaciones
	x = x + ( k1x + 2.0 * k2x + 2.0 * k3x + k4x ) / 6.0
	y = y + ( k1y + 2.0 * k2y + 2.0 * k3y + k4y ) / 6.0
        z = z + ( k1z + 2.0 * k2z + 2.0 * k3z + k4z ) / 6.0
	return x,y,z

dt = 0.002
x0 = [0.0]
y0 = [1.0]
z0 = [0.0]
t  = [0.0]
N = 20000
for n in xrange(0,N):
  # at each time step calculate new x(t),y(t),z(t)
  # and append to lists x0, y0, z0
  x,y,z = RKTwoD(x0[n],y0[n],z0[n],VDPXDot,VDPYDot,VDPZDot,dt)
  x0.append(x)
  y0.append(y)
  z0.append(z)

 t.append(t[n] + dt)
  m1 = array(x0)
  m2 = array(y0)
  m3 = array(z0)
  m4 = array(t)
  m1.tofile("putamierdax.dat", sep='\n', format = "%e")
  m2.tofile("putamierday.dat", sep='\n', format = "%e") 
  m3.tofile("putamierdaz.dat", sep='\n', format = "%e")

It works but the script time is too long and I've got to work with three files

Mark44

My understanding is that Python is an interpreted language, so if you have a loop that runs for many iterations, it's going to take a long time. The way around that is to use a language that is compiled, such as C or C++.

I don't know enough about Python to say anything intelligible about your file I/O problem...

javiergra24

Thanks for your answer Mark44

No matter if the script takes too much time (I'm doing the script also for C++ and Java), but there must be some way to save all data in a unique file. By splitting it in four I have to use Ultraedit and it's a bit tiring..

nbo10

look into numpy.savetxt(...)

story645

Thanks for your answer but it doesn't work

Sorry, I gave you the wrong args. It should be:
table = hstack([x0,y0,z0,t])

but when you open it, you need to specify the size of the array and make sure it's saved and unpacked as the right type.
you can also look into scipy.io.savemat

Quote by javiergra24

It works but the script time is too long and I've got to work with three files

I just ran your script and it took supershort, so I'm thinking the bug is that you're running this:

Code:

  m1 = array(x0)
  m2 = array(y0)
  m3 = array(z0)
  m4 = array(t)
  m1.tofile("putamierdax.dat", sep='\n', format = "%e")
  m2.tofile("putamierday.dat", sep='\n', format = "%e") 
  m3.tofile("putamierdaz.dat", sep='\n', format = "%e")

in the for loop.

javiergra24

Quote by story645

Sorry, I gave you the wrong args. It should be:
table = hstack([x0,y0,z0,t])

but when you open it, you need to specify the size of the array and make sure it's saved and unpacked as the right type.
you can also look into scipy.io.savemat

I just ran your script and it took supershort, so I'm thinking the bug is that you're running this:

Code:

  m1 = array(x0)
  m2 = array(y0)
  m3 = array(z0)
  m4 = array(t)
  m1.tofile("putamierdax.dat", sep='\n', format = "%e")
  m2.tofile("putamierday.dat", sep='\n', format = "%e") 
  m3.tofile("putamierdaz.dat", sep='\n', format = "%e")

in the for loop.

How can I run the data saving script outside the for loop?

Thank you

story645

Quote by javiergra24

How can I run the data saving script outside the for loop?

Just unindent it, so:

Code:

for n in xrange(0,N):
    x,y,z = RKTwoD(x0[n],y0[n],z0[n],VDPXDot,VDPYDot,VDPZDot,dt)
    x0.append(x)
    y0.append(y)
    z0.append(z)
    t.append(t[n] + dt)

m1 = array(x0)
m2 = array(y0)
m3 = array(z0)
m4 = array(t)
m1.tofile("putamierdax.dat", sep='\n', format = "%e")
m2.tofile("putamierday.dat", sep='\n', format = "%e") 
m3.tofile("putamierdaz.dat", sep='\n', format = "%e")

[/code]

javiergra24

Quote by story645

Just unindent it, so:

Code:

for n in xrange(0,N):
    x,y,z = RKTwoD(x0[n],y0[n],z0[n],VDPXDot,VDPYDot,VDPZDot,dt)
    x0.append(x)
    y0.append(y)
    z0.append(z)
    t.append(t[n] + dt)

m1 = array(x0)
m2 = array(y0)
m3 = array(z0)
m4 = array(t)
m1.tofile("putamierdax.dat", sep='\n', format = "%e")
m2.tofile("putamierday.dat", sep='\n', format = "%e") 
m3.tofile("putamierdaz.dat", sep='\n', format = "%e")

[/code]

Thank you very much!! So simple and I didn't notice

Similar Threads for: How to write a data file in Python
Thread	Forum	Replies
Fortran 95: WRITE to screen changes output to WRITE to file	Programming & Comp Sci	13
Write an exponential equation from this data (data table included)	Precalculus Mathematics Homework	2
Write an is_prime function in Python	Engineering, Comp Sci, & Technology Homework	1
loading file in python shell	Programming & Comp Sci	1
Loading a file (module) in Python	Computing & Technology	6

story645 Blog Entries: 3	If you keep everything as an array instead of a list, basically take x0.append(x) y0.append(y) z0.append(z) and do a vstack or hstack at the end so table = hstack(x0,y0,z0) then you just need to write out table.tofile('output.dat') numpy lists don't write out straight to dats

Mark44 Mentor	How to write a data file in Python My understanding is that Python is an interpreted language, so if you have a loop that runs for many iterations, it's going to take a long time. The way around that is to use a language that is compiled, such as C or C++. I don't know enough about Python to say anything intelligible about your file I/O problem...

javiergra24	Thanks for your answer Mark44 No matter if the script takes too much time (I'm doing the script also for C++ and Java), but there must be some way to save all data in a unique file. By splitting it in four I have to use Ultraedit and it's a bit tiring..