using target words from arrays in regex, pythons version of perls'map'

[Lance Hoffmeyer]

Hey all, in perl I was able to use the same regular expression multiple times changing one part of it
via a previously defined array and put the results into a new array

IN PERL:

my @targets = ('OVERALL RATING',
'CLOTHING', '
ITEMS',
'ACCESSORIES',
'SHOES',
'FINE JEWELRY');

my @JA13 = map {
$file2 =~/$_.*?(?:(\d{1,3}\.\d)\s+){3}/s;
} @targets;

So, in python instead of

match2 = re.search('OVEWRALL RATING.*?(?:(\d{1,3}\.\d)\s+){3} ', file2);m01 = match2.group(1) ;print m01
match2 = re.search('CLOTHING.*?(?:(\d{1,3}\.\d)\s+){3} ', file2); m02 = match2.group(1) ;print m02
match2 = re.search('ITEMS.*?(?:(\d{1,3}\.\d)\s+){3} ', file2); m03 = match2.group(1) ;print m03
match2 = re.search('ACCESSORIES.*?(?:(\d{1,3}\.\d)\s+){3} ', file2); m04 = match2.group(1) ;print m04
match2 = re.search('SHOES.*?(?:(\d{1,3}\.\d)\s+){3} ', file2); m05 = match2.group(1) ;print m05
match2 = re.search('FINE JEWELRY.*?(?:(\d{1,3}\.\d)\s+){3} ', file2); m06 = match2.group(1) ;print m06
I would have something similar to perl above:
targets = ['OVERALL RATING',
'CLOTHING', ITEMS',
'ACCESSORIES',
'SHOES',
'FINE JEWELRY']

PROPOSED CODE:

match2 = re.search(targets[i].*?(?:(\d{1,3}\.\d)\s+){3} ', file2);m[i] = match2.group(1)


Lance

[Paddy]

I don't like string interpolation within REs, it pops me out of 'RE
mode' as I scan the line.

Maybe creating a dict of matchobjects could be used in the larger
context?:
dict( [(t, re.search(t+regexp_tail, file2) for t in targets] )

(untested).

- Pad.

[John Machin]

Think about how well the above solutions scale as len(targets)
increases.

1. Make "targets" a set, not a list.
2. Have *ONE* regex which includes a bracketed match for a generic
target e.g. ([A-Z\s]+)
3. Do *ONE* regex search, not 1 per target
4. If successful, check if the bracketed gizmoid is in the set of
targets.

It's not 100% apparent whether it is possible that there can be more
than one target in the inappropriately named file2 (it is a string,
isn't it?). If so, then write your own findall-like loop wrapped
around steps 2 & 3 above. Compile the regex in advance.

HTH,
John

[John Machin]

Would you believe "steps 3 & 4"?

[Paul McGuire]

"Dennis Lee Bieber" <wl*****@ix.netcom.com> wrote in message
news:41********************************@4ax.com...

On Mon, 15 May 2006 19:41:39 -0500, Lance Hoffmeyer
<la***@augustmail.com> declaimed the following in comp.lang.python:

I would have something similar to perl above:
targets = ['OVERALL RATING',
'CLOTHING', ITEMS',
'ACCESSORIES',
'SHOES',
'FINE JEWELRY']

PROPOSED CODE:

match2 = re.search(targets[i].*?(?:(\d{1,3}\.\d)\s+){3} ', file2);m[i] = match2.group(1)
I don't do regex's, and I also don't find packing multiple statements on

one line attractive.

I concur - this kind of multiple-statements-per-line-ishness feels
gratuitous. Yes, I know they line up nicely when all 6 statements are
printed out together, but the repetition of "mNN = match2.group(1) ;print
mNN" should tell you that this might be better done with a for loop. DRY.

However... Why don't you basically do what you say you do in
Python... Substitute you targets into the expression while inside a
loop...

targets = [ "OVERALL RATING",
"CLOTHING",
"ITEMS",
"ACCESSORIES",
"SHOES",
"FINE JEWELRY" ]

results = []
for t in target:
m2 = re.search("%s.*?(?:(\d{1,3}\.\d)\s+){3}" % t, file2)
results.append(m2.group(1))
--

# by contrast, here is a reasonably Pythonic one-liner, if one-liner it must
be
results = [ re.search(r"%s.*?(?:(\d{1,3}\.\d)\s+){3}" % t, file2).group(1)
for t in targets ]

# or for improved readability (sometimes 2 lines are better than 1):
reSearchFunc = lamdba tt,ff : re.search(tt + r".*?(?:(\d{1,3}\.\d)\s+){3}",
ff).group(1)
results = [ reSearchFunc(t,file2) for t in targets ]
Resisting-the-urge-to-plug-pyparsing-ly yours,
-- Paul

[Edward Elliott]

John Machin wrote:

Would you believe "steps 3 & 4"?

How about "two pops and a pass?"

Quick! Lower the cone of silence!

--
Edward Elliott
UC Berkeley School of Law (Boalt Hall)
complangpython at eddeye dot net