ratsimp does not work when these is a root in the numerator

[homocomputeris]

from sympy import *
ratsimp(sqrt(15)/180 + S(1)/18)       

gives
Out[10]: sqrt(15)/180 + 1/18
not
(sqrt(15)+10)/180

sympy master branch at c267d55

[oscargus]

I can confirm this, but even when changing the code to return what you expects it gets simplified...

In [42]: (sqrt(15)+10)/180
Out[42]: 

[RituRajSingh878]

many types of expression are failing -

In [4]: from sympy import * 
   ...: ratsimp(log(2)/180 + S(1)/18)                                                                                                                                                                        
Out[4]: 
log(2)   1 
────── + ──
 180     18

In [5]: from sympy import * 
   ...: ratsimp(exp(2)/180 + S(1)/18)                                                                                                                                                                        
Out[5]: 
  2     
 ℯ    1 
─── + ──
180   18

In [6]: from sympy import * 
   ...: ratsimp(I/180 + S(1)/18)                                                                                                                                                                             
Out[6]: 
1     ⅈ 
── + ───
18   180

[RituRajSingh878]

ratsimp is -"Put an expression over a common denominator, cancel and reduce."
So after canceling and reducing, it may not have a common base.

[homocomputeris]

ratsimp is -"Put an expression over a common denominator, cancel and reduce."
So after canceling and reducing, it may not have a common base.

But it doesn't imply it is split back into some sum:
Reduce a Fraction: Simplify. That is, cancel out all common factors in the numerator and denominator until no common factors remain.

I believe, if ratsimp is explicitly applied, then it should always return a fraction over a common denominator, which probably should be its main purpose. Probably, expand should be the opposite of ratsimp.

[oscargus]

I also think that it ideally should keep everything with a single denominator. Not obvious where it is changed back though...

[mpjuers]

Does anyone know of a workaround for this?

[jksuom]

You could try changing return f/g in ratsimp to return Mul(f, 1/g, evaluate=False).

[RituRajSingh878]

Due to this many test case were failing.

You could try changing return f/g in ratsimp to return Mul(f, 1/g, evaluate=False).