Hi, I have a text file with 1500 words. Could it be a script that will keep the words that only have all these letters:
n i o m s c t a
If you could show me the way I would be greatful!
Can you show example input file and desired output?
Consider regular expressions:
^[niomscta]+$ This is a character class that contains just these letters made of one or more letters.
grep '^[niomscta]+$' wordfile
Something like this?
perl -lane 'print join " ", grep { /\b[niomscta]+\b/ } (split / /); ' file
/home/usr1 >cat file
sdafsa fljds dfs dsf niomscta nio msd
sad nidso scata niow nioscata
sdasda sdjf niom
miom
fdsssdsf sct
df
/home/usr1 >perl -lane 'print join " ", grep { /\b[niomscta]+\b/ } (split / /,$_); ' file
niomscta nio
scata nioscata
niom
miom
sct
Well the wordlist has words like
html:)
121212
131313
123123
654321
8675309
666666
696969
888888
1234567
21122112
12345678
asdfjkl;
hal9000
bond007
ncc1701d
ncc1701e
ncc1701
thx1138
a12345
abcd1234
What I need is to find the word that is developed by the letters "n i o m s c t a" like "monastic".I don't need more or less letters (like "maso" or "monasticasctio")
---------- Post updated at 10:35 AM ---------- Previous update was at 10:23 AM ----------
I thought something like this:
#!/bin/sh
lol=`grep '^[n]' /home/alex/Desktop/wordlist.txt`
echo $lol > /home/alex/Desktop/lol
lol1=`grep '^[i]' /home/alex/Desktop/lol`
echo $lol1 > /home/alex/Desktop/lol1
lol2=`grep '^[o]' /home/alex/Desktop/lol1`
echo $lol2 > /home/alex/Desktop/lol2
lol3=`grep '^[m]' /home/alex/Desktop/lol2`
echo $lol3 > /home/alex/Desktop/lol3
lol4=`grep '^[s]' /home/alex/Desktop/lol3`
echo $lol4 > /home/alex/Desktop/lol4
lol5=`grep '^[c]' /home/alex/Desktop/lol4`
echo $lol5 > /home/alex/Desktop/lol5
lol6=`grep '^[t]' /home/alex/Desktop/lol5`
echo $lol6 > /home/alex/Desktop/lol6
lol7=`grep '^[a]' /home/alex/Desktop/lol6`
echo "Your word is: $lol7"
I don't know if this thought is correct...Is it?
You can still use code snippet provided..However, the shortened code is below for your requirement.
perl -lane 'print if (grep {/\b[niomscta]+\b/} $_); ' file
{ perl -lane 'print if (grep {/\b[niomscta]+\b/} $); ' file } doesn't really suits me. I want to take words from the wordlist with 8 characters (niomscta=8 chars) that contain ALL the letters : n i o m s c t a
Examples:
atcsmoin
oinmatcs
matcsoin etc...
Your command { perl -lane 'print if (grep {/\b[niomscta]+\b/} $); ' file }
has this output:
alex@lol-pc:~$ perl -lane 'print if (grep {/\b[niomscta]+\b/} $_); ' /home/alex/Desktop/wordlist.txt
montana
monica
macintos
action
tomcat
cannon
tinman
nissan
station
samson
tattoo
cccccc
sonics
cosmos
mission
tintin
moomoo
that doesn't suits me at all. I want to do a kind of unscrimbling (hope I wrote it correctly) but not from all the existing words! Only from the wordlist.txt's words!
Nothing to see here 
What is this supposed to do?
alex@lol-pc:~/Desktop$ ./pip
alex@lol-pc:~/Desktop$
perl -nle'BEGIN {
@_{split / /, shift} = (1) x length($ARGV[0]);
}
keys %_ == grep $_{$_}, split // and print;
' 'n i o m s c t a' wordlist
this works perfectly!!! 10000000000000 thxes!!!!!
I misunderstood. Below one is the corrected one.
perl -lane 'print if (join("",sort (split //)) eq "acimnost");' file
The following verifies that each word contains the same amount of unique characters as are listed in letters. If that requirement is meant, then the word is filtered through tr -d. If the list of unique characters in the word matches, an empty string results and the word is printed.
#!/bin/sh
letters=niomscta
length=${#letters}
while read word; do
[ $length -ne $(echo $word | sed 's/\(.\)/\1 /g' | tr -s ' ' '\n' | sort | uniq | wc -l) ] && continue
[ -z "$(echo $word | tr -d $letters)" ] && echo $word
done < data
Determine a word's unique characters by splitting the word into one character per line, sorting, uniq'ing, reassembling into a string and then comparing to the sorted list of allowed letters.
#!/bin/sh
# letters must be sorted
letters=acimnost
while read word; do
uniqchars=$(echo $word | sed 's/\(.\)/\1 /g;y/ /\n/' | sort | uniq | paste -s - | tr -d '\t')
[ "$letters" = "$uniqchars" ] && echo $word
done < data
These solutions allow allowed letters to appear more than once. This can be easily modified by removing uniq from the pipeline.
Alister
Excellent point!
This should handle it too:
perl -nle'BEGIN {
@_{split / /, shift} = (1) x length($ARGV[0]);
}
keys %_ <= grep $_{$_}, split // and print;
' 'n i o m s c t a' infile
---------- Post updated at 07:17 PM ---------- Previous update was at 06:17 PM ----------
Based on dennis.jacob's and alister's ideas, with the almighty Z-Shell 
letters=(n i o m s c t a)
letters=${(j::)${(uo)letters}}
while read -r; do
[[ ${(j::)${(ous::)REPLY}} == $letters ]] &&
print -- $REPLY
done<wordlist
Thank you very much but I have this bug:
Let's say that in the wordlist is the word "1a2b3c"
Running your first program will have output:
alex@lol-pc:~/Desktop$ ./ty
121212
131313
123123
21122112
abcd1234
1a2b3c
cccccc
alex@lol-pc:~/Desktop$
which contains the 1a2b3c but not only this. Can you fix this?
Yes,
based on dennis.jacob's and alister's ideas again:
perl -nle'BEGIN {
$l = join "", sort grep !$_{$_}++ , split / /, shift;
}
%_ = (); $w = join "", sort grep !$_{$_}++ , split //;
print if $w eq $l;
' '1 a 2 b 3 c' wordlist
If you prefer to pass the chars without spaces:
perl -nle'BEGIN {
$l = join "", sort grep !$_{$_}++ , split //, shift;
}
%_ = (); $w = join "", sort grep !$_{$_}++ , split //;
print if $w eq $l;
' '1a2b3c' wordlist
I may have lost track of the requirements, so to not waste anyone's time I will state what the following code accomplishes: It reads one word lines and compares those words against a list of allowable letters. If the word consists of nothing but those letters, possibly occuring more than once, the word is considered a match and is printed.
As simple as I could manage without using awk or perl (I have nothing but the utmost respect for those tools, but sometimes I'm a masochist ;)):
#!/bin/sh
l=acimnost #allowed letters must be sorted
while read w; do
[ $l = "$(echo $w | sed 's/./& /g;y/ /\n/' | sort | tr -ds \\n $l)" ] && echo $w
done < data
Parameterizing it to accept the list of letters and the wordlist file as arguments is trivial and left as an exercise to whomever wishes to do it.
Cheers,
Alister
Yep,
alister's shell solution works too, the only drawback, in my opinion, is the performance (too many forks/execs per record in the pipeline).
Hi, radoulov:
Sir, how dare you doubt the performance of my code. I will have you know that my shell is faster than a speeding bullet. heheheh
Seriously, though, performance and efficiency were most definitely not a priority in that code's design. It was just a sh exercise for fun.
Cheers,
Alister
Sure, sorry 
If I have a time later, I'll try with POSIX shell too.