Remove some characters from a string

[Julien]

Hi,

I can't seem to find the right regular expression to achieve what I
want. I'd like to remove all characters from a string that are not
numbers, letters or underscores.

For example:

>>magic_functio n('si_98%u^d@.a s-*gf')

str: 'si_98udasgf'

Would you have any hint?

Thanks a lot!

Julien

[Chris]

Re: Remove some characters from a string

On Jul 17, 10:13 am, Julien <jpha...@gmail. comwrote:

Hi,
>
I can't seem to find the right regular expression to achieve what I
want. I'd like to remove all characters from a string that are not
numbers, letters or underscores.
>
For example:
>

>magic_function ('si_98%u^d@.as-*gf')

>
str: 'si_98udasgf'
>
Would you have any hint?
>
Thanks a lot!
>
Julien

One quick and dirty way would be...

import string
safe_chars = string.ascii_le tters + string.digits + '_'
test_string = 'si_98%u^d@.as-*gf'
''.join([char if char in safe_chars else '' for char in test_string])

you could also use a translation table, see string.translat e (the
table it uses can be made with string.maketran s)

[Paul Hankin]

Re: Remove some characters from a string

On Jul 17, 9:13 am, Julien <jpha...@gmail. comwrote:

Hi,
>
I can't seem to find the right regular expression to achieve what I
want. I'd like to remove all characters from a string that are not
numbers, letters or underscores.
>
For example:
>

>magic_function ('si_98%u^d@.as-*gf')

>
str: 'si_98udasgf'

For speed, you can use 'string.transla te', but simplest is to use a
comprehension:

import string

def magic_function( s, keep=string.asc ii_letters + string.digits +
'_'):
return ''.join(c for c in s if c in keep)

--
Paul Hankin

[Fredrik Lundh]

Re: Remove some characters from a string

Julien wrote:

I can't seem to find the right regular expression to achieve what I
want. I'd like to remove all characters from a string that are not
numbers, letters or underscores.
>
For example:
>

>>>magic_functi on('si_98%u^d@. as-*gf')

str: 'si_98udasgf'

the easiest way is to replace the things you don't want with an empty
string:

>>re.sub("\W" , "", "si_98%u^d@ .as-*gf")

'si_98udasgf'

("\W" matches everything that is "not numbers, letters, or underscores",
where the alphabet defaults to ASCII. to include non-ASCII letters, add
"(?u)" in front of the expression, and pass in a Unicode string).

</F>