Hi,
I am trying to write a script that will take 2 or more instances of repetitive alphabets (ZZ) to be removed from a field. This should only happen from beginning and end of a field.
For Example :
Input File
a) ZZZIBM Corporation
b) ZZZIBM Corporation ZZZZZ
b) IBM ZZZ Corporation
Output Result should be as follow :
a) IBM Corporation
b) IBM Corporation
b) IBM ZZZ Corporation
Please advise.
Thanks....
Always be ZZ?
sed 's/\(^.. \)Z*/\1/;s/Z*$//' urfile
sed "s/^\(.\)\1\{1,\}//;s/\(.\)\1\{1,\}$//" file
It will be all Z's but it can be 2 or more repetitive Z. So will the sed command will work for 2 or more Z? Also, will it take Z's from beginning and end of a field? Z should not be taken away in between the words. Please advise.
Thanks...
yes, if you try it.
I tried it and it returns the same result without cleaning ZZ.
sed 's/\(^.. \)Z*/\1/;s/Z*$//' zzz_test.dat
ZZZIBM Corporation
ZZZIBM Corporation ZZZZZ
IBM ZZZ Corporation
I also tried
sed "s/^\(.\)\1\{1,\}//;s/\(.\)\1\{1,\}$//" file
and it also not work.
Also, I need this to be done on a field not a file. I am extracting a field from a file already while looping through each line. Please advise. I am doing this in Linux OS.
Thanks...
It is working for me.
$ cat file
ZZZIBM Corporation
ZZZIBM Corporation ZZZZZ
IBM ZZZ Corporation
$ sed "s/^\(.\)\1\{1,\}//;s/\(.\)\1\{1,\}$//" file
IBM Corporation
IBM Corporation
IBM ZZZ Corporation
What is your output of sed command?
$ cat zzz_test.dat
ZZZIBM Corporation
ZZZIBM Corporation ZZZZZ
IBM ZZZ Corporation
$ sed "s/^\(.\)\1\{1,\}//;s/\(.\)\1\{1,\}$//" zzz_test.dat
IBM Corporation
IBM Corporation ZZZZZ
IBM ZZZ Corporation
It is not able to clear up end ZZZZZ. Please advise.
Thanks..
The solutions of anbu23 and rdcwayx work for me. Here's one with awk:
awk '{sub(/^Z+/,"");sub(/ Z+$/,"")}1' file
Maybe it is Linux that is causing issue.
$ awk '{sub(/^Z+/,"");sub(/ Z+$/,"")}1' zzz_test.dat
IBM Corporation
IBM Corporation ZZZZZ
IBM ZZZZ Corporation
Please advise.
Thanks....
May be because of space at the end of the line. Try this
sed "s/^\(.\)\1\{1,\}//;s/\(.\)\1\{1,\} *$//" zzz_test.dat
It is end of line character at end. No Space. Everything else is working fine except the last ZZZ.
$ sed "s/^\(.\)\1\{1,\}//;s/\(.\)\1\{1,\} *$//" zzz_test.dat
IBM Corporation
IBM Corporation ZZZZZ
IBM ZZZZ Corporation
Please advise. Thanks a bunch for the help in this. The reason for this is we are writing a cleansing routine at work to clean data. All other code is done except the last part which is where I am stuck.
Thanks again.....
Can you show the output of this command?
cat -vet zzz_test.dat
Ah...It is having ^M at end of record.
cat -ver zzz_test.dat
ZZZZIBM Corporation ^M$
ZZZZIBM Corporation ZZZZZ^M$
IBM ZZZZ Corporation ^M$
I will replace ^M and try the sed again. Thanks a lot for the help.
tr -d \\015 < zzz_test.dat | sed "s/^\(.\)\1\{1,\}//;s/\(.\)\1\{1,\} *$//"
Yeap this worked. Thanks a lot anbu23 and all others for your help.
The below code is changing all alphabets that begin with 2 or more letters in begiining to clear out.
tr -d \\015 < zzz_test.dat | sed "s/^\(.\)\1\{1,\}//;s/\(.\)\1\{1,\} *$//"
Before : HHS/CDC
After : S/CDC
Before : 55304
After : 304
I just need 2 or more ZZ to clear out at beginning and end of a field. Please advise.
Thanks....
tr -d \\015 < zzz_test.dat | sed "s/^Z\{2,\}//;s/Z\{2,\} *$//"
This worked. Thanks.....
If your awk/nawk supports multi-char field separators
awk -F'[Z ]+' '{$1=$1}1'