Programatically detecting a change in letters

hacker has asked for the wisdom of the Perl Monks concerning the following question:

I'm building a large dictionary of terms, MySQL-backed. Currently there are 3,223 entries in the database. I'm converting this into mobile format for handheld devices (constrained screen real-estate, limited "browser" capabilities, etc.)

Due to the limitations of the screen size, I've been asked to create named anchors that will go from the top of the list of terms, down to the next closest match of letters.

For example:

   <a href="#Ba">Ba</a> 
   <a href="#Be">Be</a>
   <a href="#Bi">Bi</a> 
   <a href="#Bo">Bo</a>
   <a href="#Bu">Bu</a>
[download]

Each term displayed, would lack a <a name=".."> attribute unless its letter changed from the previous term's first two letters.

The first word to begin with "Be" would have an anchor that would put the user there when they selected it from the top of the page. It should look something like this:

   <a id="339" href="?term=Bay">Bay</a><br />
   <a id="340" href="?term=Bay_Tree">Bay Tree</a><br />
   <a id="341" href="?term=Bayonet">Bayonet</a><br />

   <!-- Note that the anchor changes on this first "Be" word -->
   <a id="342" href="?term=Beach" name="Be">Beach</a><br />
[download]

It all works right now, except adding the named anchors at their respective "first instance" of words that begin with $letter . [aeiou];.

How can I detect the change in the words, as I'm outputting them one by one from my results set?

Here's a sample of the words from the live db. I've added indicators below, where anchors would logically be inserted:

__DATA__
   1930:    Nails               <-- #Na here
   1931:    Naked
   1932:    Name     
   1933:    Napkin   
   1934:    Narcotics
   1935:    Narrow
   1936:    Nature
   1937:    Nausea
   1938:    Navel
   1939:    Navy
   1940:    Nazi
   1941:    Nearsighted         <-- #Ne here
   1942:    Neck
   1943:    Necklace
   1944:    Necktie
   1945:    Necromancer
   1946:    Need
   1947:    Needle   
   1948:    Negligee
   1949:    Neighbour
   1950:    Neighbourhood
   1951:    Nephew
   1952:    Neptune
   1953:    Nerd
   1954:    Nervous_Breakdown
   1955:    Nest     
   1956:    Net      
   1957:    Nettles  
   1958:    New   
   1962:    New_Year
   1959:    News  
   1960:    Newspaper
   1961:    Newspaper_Reporter
   1963:    Nickname            <-- #Ni here
   1964:    Niece
   1965:    Night
   1966:    Nightclub
   1967:    Nightgown
   1969:    Nightingale
   1968:    Nightmare
   1970:    Ninepins 
   1971:    Nipples  
   1972:    Nobility            <-- #No here
   1973:    Noise  
   1974:    Noodles
   1975:    Noose  
   1976:    North
   1977:    Northern_Lights  
   1978:    Nose
   1979:    Notary
   1980:    Notebook
   1981:    November
   1982:    Nuclear_Bomb        <-- #Nu here
   1984:    Numbers
   1983:    Numbness 
   1985:    Nuns   
   1986:    Nuptial
   1987:    Nurse
   1988:    Nursing
   1989:    Nuts 
   1990:    Nymph
[download]

Thanks, my fellow monks.

Comment on Programatically detecting a change in letters Select or Download Code

Replies are listed 'Best First'.
Re: Programatically detecting a change in letters by Corion (Patriarch) on Sep 03, 2007 at 19:22 UTC
Why not simply look at the first two letters and see if they changed? my $last = ""; while (<DATA>) { my (undef,$number,$item,$rest) = split /\s+/, $_, 4; my $current_two_letters = substr($item,0,2); if ($current_two_letters ne $last) { print "--- Break here (change from $last to $current_two_lette +rs)\n"; }; print $_; $last = $current_two_letters; }; __DATA__ 1930: Nails <-- #Na here 1931: Naked 1941: Nearsighted <-- #Ne here 1942: Neck 1961: Newspaper_Reporter 1963: Nickname <-- #Ni here 1964: Niece 1971: Nipples 1972: Nobility <-- #No here 1973: Noise 1981: November 1982: Nuclear_Bomb <-- #Nu here 1984: Numbers 1989: Nuts 1990: Nymph [download]	[reply] [d/l]
Re^2: Programatically detecting a change in letters by Anonymous Monk on Sep 03, 2007 at 23:24 UTC
how does this approach handle the following sequence? `Aardvark Abacus Actuary Additive Aeolian` [download]	[reply] [d/l]
Re^3: Programatically detecting a change in letters by Corion (Patriarch) on Sep 04, 2007 at 06:11 UTC
Just as I expect it. The convenient thing about `__DATA__` sections is that you can put your own data into them and try stuff for yourself.	[reply] [d/l]
Re: Programatically detecting a change in letters by lodin (Hermit) on Sep 03, 2007 at 19:44 UTC
Here's a somewhat different approach that works if you only want to anchor at just `/^N[aeiou]/`. my %anchors = map { $_ => 1 } qw/ Na Ne Ni No Nu /; while (my $line = <DATA>) { my ($number, $item) = split ' ', $line; my ($first_two) = $item =~ /^(..)/ or die "Word too short: $line"; if (delete $anchors{$first_two}) { print "Insert '$first_two' anchor here!\n"; } print $line; } __DATA__ 1930: Nails <-- #Na here 1939: Navy 1940: Nazi 1941: Nearsighted <-- #Ne here 1942: Neck 1961: Newspaper_Reporter 1963: Nickname <-- #Ni here 1964: Niece 1971: Nipples 1972: Nobility <-- #No here 1973: Noise 1981: November 1982: Nuclear_Bomb <-- #Nu here 1984: Numbers [download] Result: `Insert 'Na' anchor here! 1930: Nails <-- #Na here 1939: Navy 1940: Nazi Insert 'Ne' anchor here! 1941: Nearsighted <-- #Ne here 1942: Neck 1961: Newspaper_Reporter Insert 'Ni' anchor here! 1963: Nickname <-- #Ni here 1964: Niece 1971: Nipples Insert 'No' anchor here! 1972: Nobility <-- #No here 1973: Noise 1981: November Insert 'Nu' anchor here! 1982: Nuclear_Bomb <-- #Nu here 1984: Numbers` [download] lodin	[reply] [d/l] [select]
Re: Programatically detecting a change in letters by Anonymous Monk on Sep 03, 2007 at 23:36 UTC
how about this approach: # ASSUME: words are sorted lexically ascending. use warnings; use strict; my $last_anchor = qr{ \z \A }xms; # init to never-matching pattern my $first = qr{ [A-Za-z] }xms; my $anchor = qr{ $first [AEIOUaeiou] }xms; my $letters = qr{ [_A-Za-z] }xms; my $sequence_number_field = qr{ \d+ : [ ]{4} }xms; WORD: while (<DATA>) { next WORD unless m{ $sequence_number_field (?! $last_anchor) # NO match to last anchor ($anchor $letters*) # capture anchoring word to $1 }xms; my $word = $1; $last_anchor = qr{ @{[ substr $word, 0, 2 ]} }xms; print "$word: $_"; } __DATA__ 1: xxAardvark <-- first if really Aardvark 2: Abacus 3: Actuary 4: Additive 5: Aeolian <-- first if no Aardvark 1930: Nails <-- #Na here 1931: Naked 1932: Name 1933: Napkin 1934: Narcotics 1935: Narrow 1936: Nature 1937: Nausea 1938: Navel 1939: Navy 1940: Nazi 1941: Nearsighted <-- #Ne here 1942: Neck 1943: Necklace 1944: Necktie 1945: Necromancer 1946: Need 1947: Needle 1948: Negligee 1949: Neighbour 1950: Neighbourhood 1951: Nephew 1952: Neptune 1953: Nerd 1954: Nervous_Breakdown 1955: Nest 1956: Net 1957: Nettles 1958: New 1962: New_Year 1959: News 1960: Newspaper 1961: Newspaper_Reporter 1963: Nickname <-- #Ni here 1964: Niece 1965: Night 1966: Nightclub 1967: Nightgown 1969: Nightingale 1968: Nightmare 1970: Ninepins 1971: Nipples 1972: Nobility <-- #No here 1973: Noise 1974: Noodles 1975: Noose 1976: North 1977: Northern_Lights 1978: Nose 1979: Notary 1980: Notebook 1981: November 1982: Nuclear_Bomb <-- #Nu here 1984: Numbers 1983: Numbness 1985: Nuns 1986: Nuptial 1987: Nurse 1988: Nursing 1989: Nuts 1990: Nymph [download]	[reply] [d/l]

Back to Seekers of Perl Wisdom