newby question: Splitting a string - separator

[Thomas Liesner]

Hi all,

i am having a textfile which contains a single string with names.
I want to split this string into its records an put them into a list.
In "normal" cases i would do something like:

#!/usr/bin/python
inp = open("file")
data = inp.read()
names = data.split()
inp.close()

The problem is, that the names contain spaces an the records are also
just seprarated by spaces. The only thing i can rely on, ist that the
recordseparator is always more than a single whitespace.

I thought of something like defining the separator for split() by using
a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

TIA,
Tom

 

[Michael Spencer]

Hi all,

i am having a textfile which contains a single string with names.
I want to split this string into its records an put them into a list.
In "normal" cases i would do something like:


The problem is, that the names contain spaces an the records are also
just seprarated by spaces. The only thing i can rely on, ist that the
recordseparator is always more than a single whitespace.

I thought of something like defining the separator for split() by using
a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

TIA,
Tom

\s+ gives one or more, you need \s{2,} for two or more:

>>> import re
>>> re.split("\s{2,}","Guido van Rossum Tim Peters Thomas Liesner") ['Guido van Rossum', 'Tim Peters', 'Thomas Liesner']
>>>

Michael

 

[Noah]

...
The only thing i can rely on, ist that the
recordseparator is always more than a single whitespace.

I thought of something like defining the separator for split() by using
a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

For your split regex you could say
"\s\s+"
or
"\s{2,}"

This should work for you:
YOUR_SPLIT_LIST = re.split("\s{2,}", YOUR_STRING)

Yours,
Noah

 

[Jim]

Hi Tom,

a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

For more than one, I'd use

\s\s+

-Jim

 

[James Stroud]

Hi all,

i am having a textfile which contains a single string with names.
I want to split this string into its records an put them into a list.
In "normal" cases i would do something like:




The problem is, that the names contain spaces an the records are also
just seprarated by spaces. The only thing i can rely on, ist that the
recordseparator is always more than a single whitespace.

I thought of something like defining the separator for split() by using
a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

TIA,
Tom

The one I like best goes like this:

py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> names = [n for n in data.split() if n]
py> names
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

I think it is theoretically faster (and more pythonic) than using regexes.

James

 

[Kent Johnson]

The one I like best goes like this:

py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> names = [n for n in data.split() if n]
py> names
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

I think it is theoretically faster (and more pythonic) than using regexes.

Unfortunately it gives the wrong result.

Kent

 

[Tim Peters]

[James Stroud]

The one I like best goes like this:

py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> names = [n for n in data.split() if n]
py> names
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

I think it is theoretically faster (and more pythonic) than using regexes.

[Kent Johnson]

Unfortunately it gives the wrong result.

Still, it gets extra points for being such a pleasing example ;-)

 

[bonono]

Hi all,

i am having a textfile which contains a single string with names.
I want to split this string into its records an put them into a list.
In "normal" cases i would do something like:


The problem is, that the names contain spaces an the records are also
just seprarated by spaces. The only thing i can rely on, ist that the
recordseparator is always more than a single whitespace.

I thought of something like defining the separator for split() by using
a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

Can I just use "two space" as the seperator ?

[ x.strip() for x in data.split(" ") ]

 

[James Stroud]

The one I like best goes like this:

py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> names = [n for n in data.split() if n]
py> names
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

I think it is theoretically faster (and more pythonic) than using
regexes.

Unfortunately it gives the wrong result.

Kent

Just an example. Here is the "correct version":


names = [n for n in data.split(" ") if n]

James

 

[Michael Spencer]

Hi all,

i am having a textfile which contains a single string with names.
I want to split this string into its records an put them into a list.
In "normal" cases i would do something like:

The problem is, that the names contain spaces an the records are also
just seprarated by spaces. The only thing i can rely on, ist that the
recordseparator is always more than a single whitespace.

I thought of something like defining the separator for split() by using
a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

Can I just use "two space" as the seperator ?

[ x.strip() for x in data.split(" ") ]

If you like, but it will create dummy entries if there are more than two spaces:

>>> data = "Guido van Rossum Tim Peters Thomas Liesner"
>>> [ x.strip() for x in data.split(" ") ]

['Guido van Rossum', 'Tim Peters', '', 'Thomas Liesner']

You could add a condition to the listcomp:

>>> [name.strip() for name in data.split(" ") if name]

['Guido van Rossum', 'Tim Peters', 'Thomas Liesner']

but what if there is some other whitespace character?

>>> data = "Guido van Rossum Tim Peters \t Thomas Liesner"
>>> [name.strip() for name in data.split(" ") if name] ['Guido van Rossum', 'Tim Peters', '', 'Thomas Liesner']
>>>

perhaps a smarter condition?

>>> [name.strip() for name in data.split(" ") if name.strip(" \t")]

['Guido van Rossum', 'Tim Peters', 'Thomas Liesner']

but this is beginning to feel like hard work.


I think this is a case where it's not worth the effort to try to avoid the regexp

>>> import re
>>> re.split("\s{2,}",data) ['Guido van Rossum', 'Tim Peters', 'Thomas Liesner']
>>>

Michael

 

[Steven D'Aprano]

Hi all,

i am having a textfile which contains a single string with names.
I want to split this string into its records an put them into a list.
In "normal" cases i would do something like:




The problem is, that the names contain spaces an the records are also
just seprarated by spaces. The only thing i can rely on, ist that the
recordseparator is always more than a single whitespace.

I thought of something like defining the separator for split() by using
a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

TIA,
Tom

The one I like best goes like this:

py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> names = [n for n in data.split() if n]
py> names
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

I think it is theoretically faster (and more pythonic) than using regexes.

Yes, but the correct result would be:

['Guido van Rossum', 'Tim Peters', 'Thomas Liesner']

Your code is short, elegant but wrong.

It could also be shorter and more elegant:

# your version
py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> [n for n in data.split() if n]
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

# my version
py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> data.split()
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

The "if n" in the list comp is superfluous, and without that, the whole
list comp is unnecessary.

 

[James Stroud]

Hi all,

i am having a textfile which contains a single string with names.
I want to split this string into its records an put them into a list.
In "normal" cases i would do something like:



#!/usr/bin/python
inp = open("file")
data = inp.read()
names = data.split()
inp.close()


The problem is, that the names contain spaces an the records are also
just seprarated by spaces. The only thing i can rely on, ist that the
recordseparator is always more than a single whitespace.

I thought of something like defining the separator for split() by using
a regex for "more than one whitespace". RegEx for whitespace is \s, but
what would i use for "more than one"? \s+?

TIA,
Tom

The one I like best goes like this:

py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> names = [n for n in data.split() if n]
py> names
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

I think it is theoretically faster (and more pythonic) than using regexes.

Yes, but the correct result would be:

['Guido van Rossum', 'Tim Peters', 'Thomas Liesner']

Your code is short, elegant but wrong.

It could also be shorter and more elegant:

# your version
py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> [n for n in data.split() if n]
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

# my version
py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> data.split()
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

The "if n" in the list comp is superfluous, and without that, the whole
list comp is unnecessary.

see my post from 1 hr before this one.

 

[Tim Roberts]

The one I like best goes like this:

py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> names = [n for n in data.split() if n]
py> names
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

I think it is theoretically faster (and more pythonic) than using regexes.

But it is slower than this, which produces EXACTLY the same (incorrect)
result:

data = "Guido van Rossum Tim Peters Thomas Liesner"
names = data.split()

 

[Fredrik Lundh]

py> data = "Guido van Rossum Tim Peters Thomas Liesner"
py> names = [n for n in data.split() if n]
py> names
['Guido', 'van', 'Rossum', 'Tim', 'Peters', 'Thomas', 'Liesner']

I think it is theoretically faster (and more pythonic) than using
regexes.

Unfortunately it gives the wrong result.

Just an example. Here is the "correct version":

names = [n for n in data.split(" ") if n]

where "correct" is "still wrong", and "theoretically faster" means "slightly
slower" (at least if fix your version, and precompile the pattern).

</F>