Error handling for conditional functions

[Griffe27]

Hi,
Thanks for your package!
I've a IFEQ custom function that takes 3 parameters :
value, value_true, value_false
If value==0 then returns value_true else returns value_false.
It is used to avoid division by zero : it returns 0 if denominator is zero, otherwise the division.
Problem is that value_false is evaluated even if condition is true.
Is there a way to manage that?
Thanks!

[newville]

@Griffe27 post real code that shows a problem. This works for me:

>>> from asteval import Interpreter
>>> aeval = Interpreter()
>>> aeval("""def myfunc(a, b, c=10):
...     if c != 0:
...         return (a+b)/c
...     else:
...         return 0
... """)

>>> aeval("print(myfunc(2,12,c=1))")
14.0
>>> aeval("print(myfunc(2,12,c=0))")
0

[Griffe27]

Thanks for your answer @newville , sorry I was far from precise!
Here is an example:

from asteval import Interpreter
aeval = Interpreter()
aeval("""def ifeq(a,b,t,f):
     return t if a==b else f""")
aeval.symtable['x'] = 5
aeval.eval("ifeq(x, 0, 100, 10/x)")  # returns 2.0

aeval.symtable['x'] = 0
aeval.eval("ifeq(x, 0, 100, 10/x)")  # division by zero, I would like to return 100

I fully understand that parameters should be evaluated before calling the function, but is there a way for example to catch error when parameter is not used or something like that?

If possible I would like to avoid modifying formulas as there are many already existing.
There are also different kind of call like:
aeval.eval("ifeq(x, 0, y, x)")
aeval.eval("ifeq(x, 0, 100, ifeq(y, 0, x+z, x+y))")

If not possible one way would be to create a specific function for "division" case and passing numerator and denominator separatly.

Thank you

[newville]

@Griffe27

In

aeval.symtable['x'] = 0
aeval.eval("ifeq(x, 0, 100, 10/x)")  # division by zero, I would like to return 100

the evaluation of 10/x happens in preparing the call to ifeq. It cannot be fixed inside the function.

That is the correct behavior, matching Python:

>>> def ifeq(a,b,t,f):
...     return t if a==b else f
...
>>> x = 5
>>> ifeq(x, 0, 100, 10/x)
2.0
>>> x = 0
>>> ifeq(x, 0, 100, 10/x)
Traceback (most recent call last):
  File "<python-input-5>", line 1, in <module>
    ifeq(x, 0, 100, 10/x)
                    ~~^~
ZeroDivisionError: division by zero

but is there a way for example to catch error when parameter is not used or something like that?

How would one know what 'not used' means when preparing to call the function? But also: f is used right there in return t if a==b else f. Any lexical analysis would conclude it is used.

There is no problem with the function definition or call. Any problem is completely contained in having 10/x in the call to the function. If the whole point is to avoid dividing by zero, then you really ought to avoid dividing by zero.