UNIX for Dummies Questions & Answers

Dear people, I have an inquiry..

My need is to identify all strings marked at the beginning with unique "character1" in one file and replace those by strings marked with unique "character2" from another file. The important is to replace them strictly in order or occurence in both files.. Also, the "characters" 1 and 2 are important identifiers and must stay in their original files (only the rest of strings is replaced).

Is that possible to do such with a combination of grep and sed?

thanks a lot for your valuable help

Show a sample of each file's contents and how the final result should look like.

file1:

>word1
>word2
>word3

the identifier is '>', the 'word..' can be whatever.

file2:
Query= wordx
Query= wordy
Query= wordz

the identifier is 'Query= ', the 'word..' is not identical to the one in file1, but comes in the right order (word1~wordx).

the result:

>wordx
>wordy
>wordz

Does the first file have a space after the ">"? If so, look at the "join" command.

no, it doesn't. I will try to study the 'join' to see if it helps, thanks.

oh, I guess the problem with join is that the lines in each file are separated by other lines, whitout the identifiers and with diffeent content (didn't mention this before, sorry..). And their order must not be broken, so they cannot be sorted before using join.

So the files are:

file1:

>word1
ererererererererer
>word2
ghghghghghg
hjhjhjhjhjh
>word3
bnbnbnbnbnbnb

the identifier is '>', the 'word..' can be whatever, the lines are separated by lines with whatever content.

file2:

Query= wordA
ghghgffkjghrhrhrh
qpqpqpqpqpqpqpqq
Query= wordB
fhfhfhfhf gjgjgj hkhkhkh
tyt5757
Query= wordC
ghghdjdjdj00

the identifier is 'Query= ', the 'word..' is not identical to the one in file1, but comes in the right order (word1~wordA, word2~wordB, etc). There is a space after the 'Query=', the lines are separated by other lines with different content.

the needed result:

>wordA
ererererererererer
>wordB
ghghghghghg
hjhjhjhjhjh
>wordC
bnbnbnbnbnbnb

Maybe someone has a idea how to do this automatically into shell?

Use nawk or /usr/xpg4/bin/awk on Solaris:

Code:

% head file?
==> file1 <==
>word1
ererererererererer
>word2
ghghghghghg
hjhjhjhjhjh
>word3
bnbnbnbnbnbnb

==> file2 <==
Query= wordA
ghghgffkjghrhrhrh
qpqpqpqpqpqpqpqq
Query= wordB
fhfhfhfhf gjgjgj hkhkhkh
tyt5757
Query= wordC
ghghdjdjdj00


% awk 'NR == FNR {
  if(/^Query= /)
    _[++count1] = ">"$2
  next
  }
(/^>/ && $0 = _[++count2]) || 1
' file2 file1
>wordA
ererererererererer
>wordB
ghghghghghg
hjhjhjhjhjh
>wordC
bnbnbnbnbnbnb