Passing Operators As Functions To Use With Pandas Data Frames

I am selecting data from series on basis of threshold . >>> s = pd.Series(np.random.randn(5)) >>> s 0 -0.308855 1 -0.031073 2 0.872700 3 -0.547615 4

Solution 1:

I'm all about @cᴏʟᴅsᴘᴇᴇᴅ's answer and @Zero's linked Q&A... But here is an alternative with numexpr

importnumexprasnes[ne.evaluate('s {} {}'.format(ops[cfg['op']], cfg['threshold']))]

0-0.3088551-0.0310733-0.547615Name: A, dtype: float64

---

I reopened this question after having been closed as a dup of How to pass an operator to a python function?

The question and answers are great and I showed my appreciation with up votes.

Asking in the context of a pandas.Series opens it up to using answers that include numpy and numexpr. Whereas trying to answer the dup target with this answer would be pure nonsense.

Solution 2:

Define a dictionary of methods that can stand in for your operators.

importoperator    
d = {
         'more'  : operator.gt,
         'less'  : operator.lt,
         'equal' : operator.eq, 
         'not equal' : operator.ne
   }

Now, just index into your dictionary and apply your function parameters.

m = d[cfg['op']](s, cfg['threshold'])
m

0False1True2True3False4False
dtype: bool

s[m]

1   -0.2620542   -1.300810
dtype: float64

Here,

d[cfg['op']](s, cfg['threshold']) 

Is translated into

operator.lt(s, 0)