Using string as array indices in NumPy

Question

I'm handling large numerical arrays in python through a GUI. I'd like to expose the slicing capabilities to a textbox in a GUI, so I can easily choose part of the array that should be used for the calculation at hand.

Simple example of what I'd like to do:

arr = array([0, 10, 20, 30, 40, 50, 60, 70, 80, 90])

a = "2:4" # example string from GUI Textbox
b = "[3, 4, 5]" # example string from GUI Textbox


print arr[a] # not valid code -> what should be written here to make it work?
print arr[b] # not valid code -> what should be written here to make it work?

should output:

[20, 30]
[30, 40, 50]

I found out about the slice function, but I'd need to parse my string manually and create a slice. Is there a simpler way?

is this expected to handle only 1d arrays?

Reti43
– Reti43

2017-11-17 15:59:25 +00:00
Commented Nov 17, 2017 at 15:59 — Reti43
– Reti43, Commented Nov 17, 2017 at 15:59
No, arbitrary dimension (up to fifth at this point)

Thomas
– Thomas

2017-11-19 11:11:36 +00:00
Commented Nov 19, 2017 at 11:11 — Thomas
– Thomas, Commented Nov 19, 2017 at 11:11

Paul Panzer · Accepted Answer · 2017-11-17 15:37:22Z

2

Maybe since you are only expecting a very limited character set it is acceptable using eval this once:

if not all(c in "1234567890-[],: " for c in b): # maybe also limit the length of b?
    # tell user you couldn't parse and exit this branch
slice_ = eval(f'np.s_[{b}]')
# slice_ can now be applied to your array: arr[slice_]

edited Nov 17, 2017 at 15:37

answered Nov 17, 2017 at 15:30

Paul Panzer

53.3k3 gold badges59 silver badges103 bronze badges

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

9 Comments

Elis Byberi Over a year ago

It is important to note that if start is bigger than stop, it will return an empty array. arr[5:2] is empty and no exception is raised. That's why I was talking about security (implying validity).

Paul Panzer Over a year ago

You do realize that this is a perfectly valid and useful result? There is a reason for arr[5:2] not raising an exception.

Elis Byberi Over a year ago

The reason is that it is logically understandable. Slicing out of bounds return empty list. It may be invalid for GUI interface.

Elis Byberi Over a year ago

You do realize that f-string f'np.s_[{b}]' can not be used in Python 2.7? I was wondering how OP got this working! Hahahaha!

Elis Byberi Over a year ago

It is "go off on a tangent". However, I do not know what in particular you do not understand in my statement: arr[5:2] is empty and no exception is raised. Your question is redundant: You do realize that this is a perfectly valid and useful result? I am out! Have a nice day!

|

Kishore Devaraj · Accepted Answer · 2017-11-17 14:11:39Z

0

Try this i think it works, assuming that you will get min and max in your string.

import re
a = "2:4"
min = int(min(re.findall(r'(\d)',a)))
max = int(max(re.findall(r'(\d)',a)))
print arr[min:max]

edited Nov 17, 2017 at 14:11

answered Nov 17, 2017 at 13:53

Kishore Devaraj

1569 bronze badges

2 Comments

Thomas Over a year ago

This would only work for use case a, not b (slicing with another array)

Kishore Devaraj Over a year ago

If these two are your only inputs, then you should have a logic in place to identify whether it is array or slicing type. Then you can add a if else condition to process your data. For array type, the same code will work, but just increment your max value by one

Collectives™ on Stack Overflow

Using string as array indices in NumPy

2 Answers 2

9 Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

9 Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related