How to return index of Numpy array by search

Question

I turned a panda df to np array by
df2.to_numpy()
it returns an array like:

array([['040F', '040J', nan, nan, nan, nan, nan, nan],
       ['040F', '040J', nan, nan, nan, nan, nan, nan],
       ['040F', '040J', nan, nan, nan, nan, nan, nan],
       ['040V', '007', nan, nan, nan, nan, nan, nan],
       ['040S', '040J', '040F', nan, nan, nan, nan, nan])

Note: the array "width" is not fixed
Now I get the column number which contains "040J" by

df2.columns[np.nonzero(df2.to_numpy() == "040J")[-1]]
Index(['V2', 'V2', 'V2', 'V2'],
      dtype='object', name=0)

"V2" is the column label which contains "040J" in df2
this results skipped 4th row since it contains no "040J"

This array is OK for me, but I need to now which row does not contain "040J", how can I do it?

I wish there is a way to do it, but just fail to find a method using pandas alone and produces "a list of column# of each row's 040J" — twrandy
– twrandy, Commented Feb 8, 2021 at 23:32

Stef · Accepted Answer · 2021-02-09 07:56:26Z

2

As Ferris showed you can do it directly in pandas, but if you prefer numpy you can do it like this:

a = df2.to_numpy()
(~(a=='040J').any(1)).nonzero()

edited Feb 9, 2021 at 7:56

answered Feb 8, 2021 at 8:13

Stef

30.9k3 gold badges34 silver badges60 bronze badges

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

2 Comments

twrandy Over a year ago

thanks, this also works, but the result: (array([3], dtype=int64),) is a tuple, is there a way to retrieve the number "3"?

Stef Over a year ago

nonzero returns a tuple of arrays, one for each dimension. As we know that we have exactly one dimension we can safely use ...[0]. To get the first element of the array you'll have to wrap it in a try block in case all rows contain '040J': try: i = (~(a=='040J').any(1)).nonzero()[0][0] except: i = None

Ferris · Accepted Answer · 2021-02-08 01:34:25Z

2

# which column contains '040J'
obj = (df2 == "040J").sum()
obj[obj > 0]

# which row does not contains '040J'
obj = (df2 == "040J").sum(1)
obj[obj == 0]

answered Feb 8, 2021 at 1:34

Ferris

5,6611 gold badge18 silver badges27 bronze badges

2 Comments

twrandy Over a year ago

Thanks, this is beyond my knowledge but neat, only I don't know about the result: 4 0 dtype: int64 What is 4 0? The return is actually 0, an int64 data, how do I get the 4?

Ferris Over a year ago

that mean's 4th rows has 0 040J , you can get 4 from obj[obj == 0].index.tolist()

Collectives™ on Stack Overflow

How to return index of Numpy array by search

2 Answers 2

2 Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related