Take the 2-minute tour ×
Unix & Linux Stack Exchange is a question and answer site for users of Linux, FreeBSD and other Un*x-like operating systems.. It's 100% free, no registration required.

I am using solaris 10 and Grep options involving -f doesnt work .

I have two pipe separated files:

file1:

abc|123|BNY|apple|
cab|234|cyx|orange|
def|kumar|pki|bird|

file 2 :

abc|123|
kumar|pki|
cab|234

I would like to compare the first two columns of file2 with file1 (search through the entire contents of file1 in first two columns) if they match print the matched line of file1. Then search for the second line of file 2 and so on.

Expected Output :

abc|123|BNY|apple|
cab|234|cyx|orange|

The files I have are huge, containing about 400,000 lines. So I would like to make the execution fast.

share|improve this question
    
I removed the leading spaces from your examples, if you want it, please roll the edit back. Remember that spaces are significant, you should only have them if they exist in your actual files. –  terdon Jun 6 at 10:29
    
Try using the GNU version of grep, it's under /usr/sfw/bin/ggrep. stackoverflow.com/questions/15259882/… –  slm Jun 6 at 10:55

3 Answers 3

This is what awk was designed for:

$ awk -F'|' 'NR==FNR{c[$1$2]++;next};c[$1$2] > 0' file2 file1
abc|123|BNY|apple|
cab|234|cyx|orange|

Explanation

  • -F'|' : sets the field separator to |.
  • NR==FNR : NR is the current input line number and FNR the current file's line number. The two will be equal only while the 1st file is being read.
  • c[$1$2]++; next : if this is the 1st file, save the 1st two fields in the c array. Then, skip to the next line so that this is only applied on the 1st file.

  • c[$1$2]>0 : the else block will only be executed if this is the second file so we check whether fields 1 and 2 of this file have already been seen (c[$1$2]>0) and if they have been, we print the line. In awk, the default action is to print the line so if c[$1$2]>0 is true, the line will be printed.


Alternatively, since you tagged with Perl:

perl -e 'open(A, "file2"); while(<A>){/.+?\|[^|]+/ && $k{$&}++};
         while(<>){/.+?\|[^|]+/ && do{print if defined($k{$&})}}' file1

Explanation

The first line will open file2, read everything up to the 2nd | (.+?\|[^|]+) and save that (the $& is the result of the last match operator) in the %k hash.

The second line processes file1, uses the same regex to extract the 1st two columns and print the line if those columns are defined in the %k hash.


Both of the above approaches will need to hold the 2 first columns of file2 in memory. That shouldn't be a problem if you only have a few hundred thousand lines but if it is, you could do something like

cut -d'|' -f 1,2 file2 | while read pat; do grep "^$pat" file1; done

But that will be slower.

share|improve this answer
    
But won't this load all (the first two columns) of file2 into memory? –  Joseph R. Jun 6 at 10:44
    
@terdon: awk -F'|' 'NR==FNR{c[$1$2]++;next};c[$1$2] > 0' is shorter version. –  cuonglm Jun 6 at 10:46
    
it doesnt work .. –  user68365 Jun 6 at 10:47
    
@user68365: Does file2 have duplicate rows? –  cuonglm Jun 6 at 10:48
    
NO it doesnt have any duplicate rows –  user68365 Jun 6 at 10:52

I think

grep -Ff file2 file1

is what you're looking for. It should be efficient, but I'm not sure it will be as accurate as you want. If abc|123 (for example) is found in a line in file1 in different columns, that line will be printed as well. If you can guarantee that this will never happen, the above line should work.

share|improve this answer
    
Grep wouldnt be sufficient ,since the abc|123 may be present somewhere in th file. Moreover i am using solaris 10 and i am unable to use that grep option too. –  user68365 Jun 6 at 10:29
1  
@user68365 please clarify all this in your question. You need to tell us your OS and specify that you only want to match the first 2 columns. –  terdon Jun 6 at 10:32
$  sed 's/^/\^/' 2.txt > temp.txt ; grep 1.txt -f temp.txt
abc|123|BNY|apple|
cab|234|cyx|orange|
share|improve this answer
1  
As i have edited and mentioned in the question, the grep -f options doesnt work in my system –  user68365 Jun 6 at 10:38
    
Solaris 10 has a gnu core-utils in /usr/sfw/bin Use /usr/sfw/bin/sed and /usr/sfw/bin/grep –  mr_tron Jun 6 at 12:10

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.