Removing unwanted elements in between wanted elements in an array

Question

I have an array that can be simplified to something like this:

a=[nan, 2, 1, 1, nan, nan, 3, nan, nan, 3, 4, 7, 8, 9, 8, 7, 7, nan, nan, nan, 3, 2, nan, 2]

I would like to keep the elements 3, 4, 7, 8, 9, 8, 7, 7 the nan elements don't bother me because they don't show up in the graphic, but I would like to filter the rest of the values.

Criteria: Longest sequence without a nan

What **exactly** is the criteria for the numbers you want to keep? — martineau, Apr 27 '17 at 23:12
The criteria would be something along the lines of the longest sequence without an interruption (nan) — Ricardo Fumachi, Apr 27 '17 at 23:26
"Something along the lines"? If you know exactly what you want, edit that specification into your question. If you don't know what you want, you're not ready to ask on SO yet. — DSM, Apr 27 '17 at 23:27

Efron Licht · Answer 1 · 2017-04-27T23:25:49.303

2

This is an excellent place for a list comprehension.

The syntax is as follows: [x for x in my_list if condition()]

So, assuming your criteria are x is nan or x>=3:

from math import isnan
filtered_list = [x for x in my_list if isnan(x) or x >=3]

Note that nan behaves unusually: for any variable v, (nan==v)==False . Use math.isnan instead.

edited Apr 27 '17 at 23:25

answered Apr 27 '17 at 23:19

Efron Licht

488
3
13

Peter Wood · Accepted Answer · 2017-04-27T23:47:31.040

2

Substantially edited since the question is clarified.

To get the longest sequence you can use groupby and max:

>>> from itertools import groupby
>>> from math import isnan

>>> nan = float('NaN')

>>> a = [nan, 2, 1, 1, nan, nan, 3, nan, nan, 3, 4, 7, 8,
...      9, 8, 7, 7, nan, nan, nan, 3, 2, nan, 2]

>>> max((list(group) for key, group in groupby(a, key=isnan)), key=len)
[3, 4, 7, 8, 9, 8, 7, 7]

edited Apr 27 '17 at 23:47

answered Apr 27 '17 at 23:19

Peter Wood

23,859
5
60
99

This is wrong. NaN compares not equal with all elements as part of the IEEE specification. 'float('nan') == float('nan')` -> False 'float('nan') != float('nan') -> True – Efron Licht Apr 27 '17 at 23:20
@EfronLicht I wasn't think `nan` was the same as `NaN`. I should have clarified before answering. – Peter Wood Apr 27 '17 at 23:24
I would like to keep only the sequence '3, 4, 7, 8,9, 8, 7, 7' not the other values – Ricardo Fumachi Apr 27 '17 at 23:25
1

@RicardoFumachi Why not the other values? – Peter Wood Apr 27 '17 at 23:25
@PeterWood the question is ambiguous. I'm making an (educated) guess too. – Efron Licht Apr 27 '17 at 23:28
Respectfully, I have trouble believing that transcript is real. That's not how `groupby` works. – DSM Apr 27 '17 at 23:37
I see, I was not familiar with the groupby function, it works. – Ricardo Fumachi Apr 27 '17 at 23:45

score -1 · Answer 3 · edited May 23 '17 at 12:10

-1

You could just try

     a = a[9:-7]

after your current code...

See Colon (:) in Python list index for explanation

edited May 23 '17 at 12:10

Community

1
1

answered Apr 27 '17 at 23:14

valleymanbs

487
3
14

Yes, but I would like a way to determine the starting and ending point without having to actually look at the array. by playing with the nans maybe? – Ricardo Fumachi Apr 27 '17 at 23:19
1

What is the criteria? The longest sequence without `nan`? – Peter Wood Apr 27 '17 at 23:20
If you need an automated solution yes, but you need to be more specific with your question... how do you choose the elements? Position? Relative position to others (I.e. Nan elements)? – valleymanbs Apr 27 '17 at 23:22
Yes, the criteria is as Peter said the longest sequence without a nan – Ricardo Fumachi Apr 27 '17 at 23:27

Removing unwanted elements in between wanted elements in an array

3 Answers3