Python Numpy with Datafame np.select with Array values

Question

I'm trying to generate a new column based on multiple conditions of various columns. My code runs without traceback errors. Below is a snippet of the dataframe and code.

import pandas as pd
import numpy as np

dfc = pd.read_csv(r'C:\\Users\\...01.csv', header='infer')

condition = [dfc['N']==0, dfc['count']==dfc['N'], (dfc['count'] > dfc['N']) & (dfc['N'] != 0)]
rng_result = [str(dfc['i']) + '-' + str(dfc['a']),'None','None to Many'] 
dfc['rng'] = np.select(condition, rng_result, np.nan)

dfc.to_csv(r'C:\\Users\\...R_01.csv', index=False)

It might be that I don't understand numpy, the middle and last conditions come out fine. The first condition provide an array, which is not wanted. I want a string with the rows 'i' and 'a' value as I typed it below.

Could you please provide the input and the expected output as text in the question, so others can easily reproduce it? Also, if I'm reading the data correctly, the only columns that impact 'rng' are i and a - right? — Roy2012
– Roy2012, Commented Jul 9, 2020 at 6:48

r-beginners · Accepted Answer · 2020-07-09 07:19:11Z

2

The conditions are not readily apparent from the question, but is the question intended to address the following? I'm using np.where() to combine strings in the case of 'True'.

import pandas as pd
import numpy as np
import io

data = '''
count,i,a,N
1,1.4,1.4,0
1,0,0,1
2,110,140,0
3,0,0,3
4,3.5,5.1,0
4,19,22,0
'''

df = pd.read_csv(io.StringIO(data), sep=',')
df['rng'] = np.where((df['N'] == 1)|(df['N'] == 3), None, df['i'].astype(str)+'-'+df['a'].astype(str))
df
count   i   　　a     　　N rng
0   1   1.4 　　1.4   　　0 1.4-1.4
1   1   0.0 　　0.0   　　1 None
2   2   110.0   140.0   0   110.0-140.0
3   3   0.0 　　0.0   　　3 None
4   4   3.5 　　5.1   　　0 3.5-5.1
5   4   19.0    22.0    0   19.0-22.0

answered Jul 9, 2020 at 7:19

r-beginners

35.6k3 gold badges20 silver badges41 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Shane S Over a year ago

Thanks for the answer. Can this be modified for more conditions and choices? Like the condition df['count'] > df['N'] would result in "many".

r-beginners Over a year ago

Conditions can be added with &(and) or|(or).

Shane S Over a year ago

It does not work to have more than 2 conditions (Cons) because when I have 3 Cons I get the error. "TypeError: where() takes from 1 to 3 positional arguments but 4 were given." This is my mental picture of np.where( Con_1 | Con_2, choice_1, choice_2 ) because Con_3 and choice_3 is not allowed, but I want 3.

r-beginners Over a year ago

np.where has two choices: true='A', False='B'. You can't do more than that. It is possible to have multiple conditions like (A==1)|((B==3)&(C==2)).

Rob Raymond · Accepted Answer · 2020-07-09 07:21:38Z

1

Using apply I find it more readable and maintainable

data = [["count","i","a","N"],
       [1,1.4,1.4,0],
       [1,0,0,1],
       [2,110,140,0],
       [3,0,0,3],
        [4,3.5,5.1,0],
        [4,19,22,0],
        [50,0,17,21],
        [25,0,0,25]]

def cond(r):
    val = "tbd"
    if r.N==0: val = str(r.i)+"-"+str(r.a) 
    if r["count"]==r.N: val = "None"
    if (r['count'] > r['N']) and (r['N'] != 0): val = 'None to Many'
    return val

df = pd.DataFrame(data[1:], columns=data[0])
df["rng"] = df.apply(lambda r: cond(r), axis=1)

answered Jul 9, 2020 at 7:21

Rob Raymond

31.5k3 gold badges19 silver badges34 bronze badges

2 Comments

Shane S Over a year ago

I'm a little confused in the if statements. "i,a,N" are all columns. In the line: if r.N==0: val = str(r.i)+"-"+str(r.a) should this be corrected to: if r['N']==0: val = str(r['i'])+"-"+str(r['a']) ?

Rob Raymond Over a year ago

I've used mixed conventions for accessing columns. r.N is synonymous with r["N"]. Some code I typed, other code I copy/pasted....

Collectives™ on Stack Overflow

Python Numpy with Datafame np.select with Array values

2 Answers 2

4 Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related