list comprehension

**Aahz** · Jul 18 '05, 10:46 AM

Re: list comprehension

In article <c7mii9$988$1@l ust.ihug.co.nz> ,
Guy Robinson <guy@NOSPAM.r-e-d.co.nz> wrote:[color=blue]
>
>Trying to change a string(x,y values) such as :
>
>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>
>into (x,-y):
>
>out = "114320,-69808 114272,-69920 113568,-71600 113328,-72272"
>
>I tried this:
>
>print [(a[0],-a[1] for a in x.split(',')) for x in e]
>
>But it doesn't work. Can anyone suggest why or suggest an alternative
>way? The text strings are significantly bigger than this so performance
>is important.[/color]

When in doubt, break a problem into smaller steps. Let me play Socrates
for a minute: what's the first step in working with your dataset?
--
Aahz (aahz@pythoncra ft.com) <*> http://www.pythoncraft.com/

Adopt A Process -- stop killing all your children!

**Guy Robinson** · Jul 18 '05, 10:46 AM

Re: list comprehension

This works I was just wondering if something could be written more
concisely and hopefully faster:

s = "114320,698 08 114272,69920 113568,71600 113328,72272"
e = s.split(' ')
out =''
for d in e:
d =d.split(',')
out +='%s,%d ' %(d[0],-int(d[1]))
print out

Guy

Aahz wrote:
[color=blue]
> In article <c7mii9$988$1@l ust.ihug.co.nz> ,
> Guy Robinson <guy@NOSPAM.r-e-d.co.nz> wrote:
>[color=green]
>>Trying to change a string(x,y values) such as :
>>
>>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>>
>>into (x,-y):
>>
>>out = "114320,-69808 114272,-69920 113568,-71600 113328,-72272"
>>
>>I tried this:
>>
>>print [(a[0],-a[1] for a in x.split(',')) for x in e]
>>
>>But it doesn't work. Can anyone suggest why or suggest an alternative
>>way? The text strings are significantly bigger than this so performance
>>is important.[/color]
>
>
> When in doubt, break a problem into smaller steps. Let me play Socrates
> for a minute: what's the first step in working with your dataset?[/color]

**Andrew Bennetts** · Jul 18 '05, 10:46 AM

Re: list comprehension

On Mon, May 10, 2004 at 12:32:20PM +1200, Guy Robinson wrote:[color=blue]
> Hello,
>
> Trying to change a string(x,y values) such as :
>
> s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>
> into (x,-y):
>
> out = "114320,-69808 114272,-69920 113568,-71600 113328,-72272"
>
> I tried this:
>
> print [(a[0],-a[1] for a in x.split(',')) for x in e]
>
> But it doesn't work. Can anyone suggest why or suggest an alternative
> way? The text strings are significantly bigger than this so performance
> is important.[/color]

It has several syntax errors; (a[0],-a[1] for a in x.split(',')) is not a
valid expression because list comprehensions are bracketed by square
brackets, not parentheses. Also, the first part of a list comprehension,
the expression to calculate each element, needs to be in parens to if it has
a comma, so that the parser can disambiguate it from an ordinary list.

I also don't know where you got 'e' from. Is it 's', or 's.split()'?

If list comprenhensions grow unwieldy, just use a for loop. They're
probably easier to read than a list comprehension that takes you ten minutes
to concoct, and performance is almost identical. For the sake of answering
your question, though, here's a expression that does what you ask:
[color=blue][color=green][color=darkred]
>>> s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>>> s.replace(',',' ,-')[/color][/color][/color]
'114320,-69808 114272,-69920 113568,-71600 113328,-72272'

You could do this with list comprehensions, e.g.:
[color=blue][color=green][color=darkred]
>>> ' '.join(['%s,-%s' % tuple(x) for x in [pairs.split(',' ) for pairs in s.split(' ')]])[/color][/color][/color]
'114320,-69808 114272,-69920 113568,-71600 113328,-72272'

But I don't really see the point, given the way you've described the
problem.

-Andrew.

**Shalabh Chaturvedi** · Jul 18 '05, 10:46 AM

Re: list comprehension

Guy Robinson wrote:
[color=blue]
> This works I was just wondering if something could be written more
> concisely and hopefully faster:
> s = "114320,698 08 114272,69920 113568,71600 113328,72272"
> e = s.split(' ')
> out =''[/color]

outl = []
[color=blue]
> for d in e:
> d =d.split(',')
> out +='%s,%d ' %(d[0],-int(d[1]))[/color]

outl.append(s) # where s is the string you construct
out = ' '.join(outl)
[color=blue]
> print out
>
> Guy
>[/color]

It's faster to collect strings in a list and join them later than to
concatenate strings one by one.

Also, do you have to convert d[1] to int? If you are sure that it is
always a positive integer, you can do '%s,-%s' % (d[0],d[1]).

In fact you could even try to replace ',' with ',-' instead of splitting
the string at all. Of course it depends on what the format of your
incoming string is.

--
Shalabh

**Aahz** · Jul 18 '05, 10:46 AM

Re: list comprehension

In article <c7ml6k$c88$1@l ust.ihug.co.nz> ,
Guy Robinson <guy@NOSPAM.r-e-d.co.nz> wrote:[color=blue]
>
>This works I was just wondering if something could be written more
>concisely and hopefully faster:
>
>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>e = s.split(' ')
>out =''
>for d in e:
> d =d.split(',')
> out +='%s,%d ' %(d[0],-int(d[1]))
>print out[/color]

Performance I can understand (which Shalabh addressed quite nicely), but
why do you care about compressing the source code? This is simple,
straightforward , and easy to read; surely that's more important than
saving a few bytes?
--
Aahz (aahz@pythoncra ft.com) <*> http://www.pythoncraft.com/

Adopt A Process -- stop killing all your children!

**Tim Roberts** · Jul 18 '05, 10:46 AM

Re: list comprehension

Guy Robinson <guy@NOSPAM.r-e-d.co.nz> wrote:
[color=blue]
>This works I was just wondering if something could be written more
>concisely and hopefully faster:
>
>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>e = s.split(' ')
>out =''
>for d in e:
> d =d.split(',')
> out +='%s,%d ' %(d[0],-int(d[1]))
>print out[/color]

Well, why not the more obvious: s.replace(',',' ,-')
--
- Tim Roberts, timr@probo.com
Providenza & Boekelheide, Inc.

**Peter Otten** · Jul 18 '05, 10:46 AM

Re: list comprehension

Tim Roberts wrote:
[color=blue]
> Guy Robinson <guy@NOSPAM.r-e-d.co.nz> wrote:
>[color=green]
>>This works I was just wondering if something could be written more
>>concisely and hopefully faster:
>>
>>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>>e = s.split(' ')
>>out =''
>>for d in e:
>> d =d.split(',')
>> out +='%s,%d ' %(d[0],-int(d[1]))
>>print out[/color]
>
> Well, why not the more obvious: s.replace(',',' ,-')[/color]

But beware negative numbers:
[color=blue][color=green][color=darkred]
>>> int("--1")[/color][/color][/color]
Traceback (most recent call last):
File "<stdin>", line 1, in ?
ValueError: invalid literal for int(): --1[color=blue][color=green][color=darkred]
>>>[/color][/color][/color]

Therefore at least
[color=blue][color=green][color=darkred]
>>> "123,234 567,-789".replace(", ", ",-").replace( "--", "")[/color][/color][/color]
'123,-234 567,789'[color=blue][color=green][color=darkred]
>>>[/color][/color][/color]

As to robustness, the OP relying on commas not being followed by a space
seems dangerous, too, if the original data is created manually.

Peter

**Ryan Paul** · Jul 18 '05, 10:46 AM

Re: list comprehension

On Mon, 10 May 2004 12:32:20 +1200, Guy Robinson wrote:
[color=blue]
> Hello,
>
> Trying to change a string(x,y values) such as :
>
> s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>
> into (x,-y):
>
> out = "114320,-69808 114272,-69920 113568,-71600 113328,-72272"
>
> I tried this:
>
> print [(a[0],-a[1] for a in x.split(',')) for x in e]
>
> But it doesn't work. Can anyone suggest why or suggest an alternative
> way? The text strings are significantly bigger than this so performance
> is important.
>
> TIA,
>
> Guy[/color]

this works:

s = "114320,698 08 114272,69920 113568,71600 113328,72272"

o = [(int(a[0]),-int(a[1])) for a in [b.split(',') for b in s.split(' ')]]
print o

[(114320, -69808), (114272, -69920), (113568, -71600), (113328, -72272)]

**Guy Robinson** · Jul 18 '05, 10:46 AM

Re: list comprehension

s.replace(',',' ,-') HA!!:-)

Thanks Andrew. As usual making it more complicated than it needs to be...

Guy

Andrew Bennetts wrote:[color=blue]
> On Mon, May 10, 2004 at 12:32:20PM +1200, Guy Robinson wrote:
>[color=green]
>>Hello,
>>
>>Trying to change a string(x,y values) such as :
>>
>>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>>
>>into (x,-y):
>>
>>out = "114320,-69808 114272,-69920 113568,-71600 113328,-72272"
>>
>>I tried this:
>>
>>print [(a[0],-a[1] for a in x.split(',')) for x in e]
>>
>>But it doesn't work. Can anyone suggest why or suggest an alternative
>>way? The text strings are significantly bigger than this so performance
>>is important.[/color]
>
>
> It has several syntax errors; (a[0],-a[1] for a in x.split(',')) is not a
> valid expression because list comprehensions are bracketed by square
> brackets, not parentheses. Also, the first part of a list comprehension,
> the expression to calculate each element, needs to be in parens to if it has
> a comma, so that the parser can disambiguate it from an ordinary list.
>
> I also don't know where you got 'e' from. Is it 's', or 's.split()'?
>
> If list comprenhensions grow unwieldy, just use a for loop. They're
> probably easier to read than a list comprehension that takes you ten minutes
> to concoct, and performance is almost identical. For the sake of answering
> your question, though, here's a expression that does what you ask:
>
>[color=green][color=darkred]
>>>>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>>>>s.replace(' ,',',-')[/color][/color]
>
> '114320,-69808 114272,-69920 113568,-71600 113328,-72272'
>
> You could do this with list comprehensions, e.g.:
>
>[color=green][color=darkred]
>>>>' '.join(['%s,-%s' % tuple(x) for x in [pairs.split(',' ) for pairs in s.split(' ')]])[/color][/color]
>
> '114320,-69808 114272,-69920 113568,-71600 113328,-72272'
>
> But I don't really see the point, given the way you've described the
> problem.
>
> -Andrew.
>
>[/color]

**David MacQuigg** · Jul 18 '05, 10:52 AM

Re: list comprehension

On Mon, 10 May 2004 12:32:20 +1200, Guy Robinson
<guy@NOSPAM.r-e-d.co.nz> wrote:
[color=blue]
>Hello,
>
>Trying to change a string(x,y values) such as :
>
>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>
>into (x,-y):
>
>out = "114320,-69808 114272,-69920 113568,-71600 113328,-72272"
>
>I tried this:
>
>print [(a[0],-a[1] for a in x.split(',')) for x in e]
>
>But it doesn't work. Can anyone suggest why or suggest an alternative
>way? The text strings are significantly bigger than this so performance
>is important.[/color]

Seems like this is a very common problem, needing to process a
substring within a long list of strings. I like the way Ruby handles
these problems. Maybe a future version of Python could do this:

print s.split(' ').map().split( ',').reflecty() .join(' ')

You would need to define the reflecty() function, but the others
should be standard. Depending on how much variation you expect in the
input data, reflecty() could do whatever checking is necessry to avoid
the problems other posters have mentioned. Assuming the inputs are
valid string representations of numbers (i.e. no double minuses,
etc.), a simple definition could be:

def reflecty():
x,y = __self__ # a two-string list
if y[0] == '-':
return [ x, y[1:] ]
else:
return [ x, '-' + y ]

The above syntax is neither Ruby nor Python, but the idea of handling
complex sequences of string operations step-by-step, left-to-right was
inspired by Ruby. See http://userlinux.com/cgi-bin/wiki.pl?RubyPython
for a comparison of Ruby and Python string operations.

-- Dave

**Christopher Barker** · Jul 18 '05, 10:53 AM

Re: list comprehension

David MacQuigg wrote:[color=blue]
> <guy@NOSPAM.r-e-d.co.nz> wrote:[color=green]
>>Trying to change a string(x,y values) such as :
>>
>>s = "114320,698 08 114272,69920 113568,71600 113328,72272"
>>
>>into (x,-y):
>>
>>out = "114320,-69808 114272,-69920 113568,-71600 113328,-72272"[/color][/color]

if you can count on there being no spaces between the x,y pairs, this works:

" ".join([ "%s,%i"%(y[0], -int(y[1]) ) for y in [x.split(",") for x in
s.split()] ])

Though I don't think I'd do it as a one liner.

-Chris

--
Christopher Barker, Ph.D.
Oceanographer

NOAA/OR&R/HAZMAT (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception

Chris.Barker@no aa.gov

list comprehension

list comprehension

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment