The python script below evaluates the first term in the expression, 4.939e-3xAxB+8.7989 at (A,B) = (1.0,1.0):
import sympy
from sympy import *
A = sympy.Symbol(A)
B = sympy.Symbol(B)
F = 4.939e-3*A*B+8.7989
G = str(F).split("+")[0]
H = lambdify([A,B], G, "numpy")
print H(1,1)
The ouput is: 0.004939
Now the code below aims to achieve the same objective:
A = sympy.Symbol(A)
B = sympy.Symbol(B)
F = 4.939e-3*A*B+8.7989
G = str(F).split("+")[0]
H = lambdify(A, G, "numpy")
I = lambdify(B, H(1), "numpy")
print I(1)
But this returns the error:
NameError: global name 'B' is not defined
Can someone kindly explain this?
You are conflating strings and symbols at multiple instances. The only reason that any of this works at all is that SymPy applies sympify to many inputs (sympify("A")==Symbol("A")) and Python’s duck typing. Specifically, G is a string, when it should be a symbolic expression, and all first arguments passed to lambdify are strings or lists of strings, when they should be symbols or lists thereof.
A clean version of your first code would be:
from sympy.abc import A,B
import sympy
G = 4.939e-3*A*B # now this is a symbolic expression, not a string
H = lambdify( [A,B], G, "numpy" )
print(H(1,1))
And the same for your second code:
from sympy.abc import A,B
import sympy
G = 4.939e-3*A*B # now this is a symbolic expression, not a string
H = lambdify( A, G, "numpy" )
I = lambdify( B, H(1), "numpy" )
print(I(1))