in reply to Golf challange: match U.S. State names

A hybrid approach, for 110 109 (including the honorary state KT). 
print

/^\w\w\z/&("MNCAKSCOHINMOKTNVALARIDCTX SDE GAZ ORIL"=~/$_/|
/(FL|IA|KY|M[ADEIST]|N[DEHNJY]|PA|[UV]T|W[AIVY])/)

for qw(AK AL AR AZ CA CO CT DC DE FL GA HI IA ID IL IN KS
       KT KY LA MA MD ME MI MN MO MS MT NC ND NE NH NJ NM
       NV NY OH OK OR PA RI SC SD TN TX UT VA VT WA WI WV
       WY)

[download]
Getting here involved finding the longest unbroken string that could hold overlapping state names, adding a few more with breaks (spaces), then handling the strays with a regexp. A close inspection will show that Rhode Island is handled twice. That actually saved strokes.

I started down the path of looking for the shortest string that contains all of the state names (with breaks). That script is still running... So far, the solution above is shorter.

Replies are listed 'Best First'.
Re: Re: Golf challange: match U.S. State names
by dws (Chancellor) on Jun 06, 2001 at 22:13 UTC
    Now 108 
    /^\w\w\z/&("MNCAKSCOHINMOKTNVALARIDCTX SDE GAZ ORIL IAKY"=~/$_/|
/(FL|M[ADEIST]|N[DEHNJY]|PA|[UV]T|W[AIVY])/)

    [download]
    Another stoke off by adding a redundant check for Alaska, which allowed me to move Iowa and Kentucky out of the regexp and into the (more compact) string.

    This suggests that there's a lesson in here about how a bit of redundancy in data can lead to more compact representations. I think we often hear problems like "match every state and nothing else" as "match every state exactly once, and nothing else."

[reply]
[d/l]
Re: Re: Golf challenge: match U.S. State names
by petral (Curate) on Jun 06, 2001 at 16:21 UTC
    Or 105:
    print

/^..\z/&("MNCAKSCOHINMOKTNVALARIDCTX
SDE
GAZ
ORIL"=~/$_/|
/FL|IA|KY|M[ADEIST]|N[DEHJY]|PA|[UV]T|W[AIVY]/)

for qw(AK AL AR AZ CA CO CT DC DE FL GA HI IA ID IL IN KS
       KT KY LA MA MD ME MI MN MO MS MT NC ND NE NH NJ NM
       NV NY OH OK OR PA RI SC SD TN TX UT VA VT WA WI WV
       WY MY MNC .* ), " SD", "AZ\n"

    [download]
    (Dropping the inner parens around the or-chain and disregarding the state of NN which is not being honored today.)

    OOPSDATE: Yep, matches if $_='.*' (see runriq immediately below and slightly to the right).  It needs /\Q$_/, as so often happens, which defeats the character savings.  I was looking at the \w as avoiding matching the spaces (which it also does).    (Oh, and thanks runriq and I'm sorry too!)
      p

[reply]
[d/l]
[select]
      Sorry, but your expression makes '.*' an honorary state :-)

      That's why dws uses '\w\w' instead of '..' in his answer.

[reply]

In Section Meditations