// phonecodes

/ map characters to integers, setup

setup:{D@:p:=C::#:'D::digit'W;W@:p;C?:;P::=:'D;H::_hash'D::D@'P[;;0]}
digit:{M[1;M[0]?/:_dv[(_ci@[!256;65+!26;+;32])@_ic x;"\""]]}
M:("ejnqrwxdsyftamcivbkulopghz";"01112223334455666777888999")

/ phonecodes 

codes:{x,/:match[1;":";x _dvl"/-"]}
match:{[b;m;n]:[~#n;,m;#r:,/find[m;n]'[D;W;C;P;H];r;b;_f[0;m," ",*n;1_ n];r]}
find:{[m;n;d;w;c;p;h]:[c>#n;();(#d)>i:(d;h)?c#n;,/w[p i]{m," ",x,y}/:\:match[1;"";c _ n];()]}

/ test phone numbers

N:("112"
   "5624-82"
   "4824"
   "0721/608-4067"
   "10/783--5"
   "1078-913-5"
   "381482"
   "04824")

/ word list

W:("an"
   "blau"
   "Bo\""
   "Boot"
   "bo\"s"
   "da"
   "Fee"
   "fern"
   "Fest"
   "fort"
   "je"
   "jemand"
   "mir"
   "Mix"
   "Mixer"
   "Name"
   "neu"
   "o\"d"
   "Ort"
   "so"
   "Tor"
   "Torf"
   "Wasser")

/ run

setup[]

`0:,/codes'N

\

/ w1000 data (1000 phone numbers, 75000 words), 2.2 seconds 300mhz pentium 2

\cd phone
W:0:"w.txt"
N:0:"n.txt"
O:0:"o.txt"
setup[]
\t R:,/codes'N

\

/ test results

5624-82: mir Tor
5624-82: Mix Tor
4824: fort
4824: Tor 4
4824: Torf
10/783--5: je Bo" da
10/783--5: je bo"s 5
10/783--5: neu o"d 5
381482: so 1 Tor
04824: 0 fort
04824: 0 Tor 4
04824: 0 Torf

/ slow version #1, uses _sm:

M:{@[_n;`$',/x;:;&#:'x]}("e";"jnq";"rwx";"dsy";"ft";"am";"civ";"bku";"lop";"ghz")
map:{1!"*",,/$M@`$'x}
lc:(_ci@[!256;65+!26;+;32])@_ic
nq:_dv[;"\""]
D:(map nq lc@)'W
codes:{x,/:match[1;":";x _dvl"/-"]}
match:{[b;p;n]:[~#n;,p;#r:,/{match[1;p," ",W x;(-1+#D x)_ n]}'initial n;r;b;match[0;p," ",*n;1_ n];r]}
initial:{[n]k@&n _sm/:D k:&D[;0]=*n}

// slow version #2, uses _ss

M:{@[_n;`$',/x;:;&#:'x]}("e";"jnq";"rwx";"dsy";"ft";"am";"civ";"bku";"lop";"ghz")
map:{,/$M@`$'x}	
lc:(_ci@[!256;65+!26;+;32])@_ic
nq:_dv[;"\""]
C:#:'D:(map nq lc@)'W

codes:{
 o@:i:&0<#:'o:n _ss/:D k:&~C>#n:x _dvl"/-"
 i@:&#:'o;i@:j:<o:,/o;o@:j
 x,/:match[n;W k i;o;o+#:'D k i;1;":";0]}

match:{[n;w;o;k;b;p;i]:[~i<#n;,p;#r:,/{match[n;w;o;k;1;p," ",w x;k x]}'&o=i;r;b;match[n;w;o;k;0;p," ",n i;1+i];r]}

// buggy version #3, uses hashed ?

M:{@[_n;`$',/x;:;&#:'x]}("e";"jnq";"rwx";"dsy";"ft";"am";"civ";"bku";"lop";"ghz")
setup:{H::_hash'D@:p:=C::#:'D::{,/$M@`$'_dv[(_ci@[!256;65+!26;+;32])@_ic x;"\""]}'W;W@:p;C?:}
codes:{x,/:match[1;":";x _dvl"/-"]}
match:{[b;p;n]:[~#n;,p;#r:,/find[p;n]'[C;H;D;W];r;b;_f[0;p," ",*n;1_ n];r]}
find:{[p;n;c;h;d;w]:[c>#n;();(#d)>i:(d;h)?c#n;match[1;p," ",w i;c _ n];()]}

// slow version #4, uses ~\:

M:{@[_n;`$',/x;:;&#:'x]}("e";"jnq";"rwx";"dsy";"ft";"am";"civ";"bku";"lop";"ghz")
setup:{D@:p:=C::#:'D::{,/$M@`$'_dv[(_ci@[!256;65+!26;+;32])@_ic x;"\""]}'W;W@:p;C?:}
codes:{x,/:match[1;":";x _dvl"/-"]}
match:{[b;p;n]:[~#n;,p;#r:,/find[p;n]'[C;D;W];r;b;_f[0;p," ",*n;1_ n];r]}
find:{[p;n;c;d;w]:[c>#n;();#i:&d~\:c#n;,/{match[1;p," ",x;c _ n]}'w i;()]}