Vfrdtyky

Read line in file1, if line not exist in file2 write this line to the new file .

Comparison is made by the length of the hash string .

Hash values ​​in file1:

cf03189f5b05eb1a9658f80d7a0e9f02:_#.g}

edbe6de8b3ee19b45e092147f57af7b8:]mNon

47253940f843f258ffd265d13f365d70:/u'yv

5701aa8e2aa7e1cfd16ca4076bd1732a:@AQ1z

b3c0866e6fd56776bc4a18d3c87cc725:t$5OV

7a1e6090568e076c55df9dc7abf356c6:9rC@p

04046da33706518d9b15a38bcddb448e:!DFPk

Hash values ​​in file2:

edbe6de8b3ee19b45e092147f57af7b8:]mNon:str1

47253940f843f258ffd265d13f365d70:/u'yv:2str

b3c0866e6fd56776bc4a18d3c87cc725:t$5OV:3str1ng

This is a working example written in C:

#include <stdio.h>

#include <string.h>



#define LINE_LENGTH 80

typedef enum {FALSE=0,TRUE=1} BOOL;

int main(int argc, char *argv)

{

    FILE   *fpin1 = NULL;

    FILE   *fpin2 = NULL;

    FILE   *fpout = NULL;

    char   line [LINE_LENGTH]={0};

    char   line1[LINE_LENGTH]={0};

    BOOL   bCheck = FALSE;

    size_t ncHash = 0;

    size_t count = 0;



    if (argc != 4)

    {

        printf("Usage:%s <file1> <file2> <OutFile>n", argv[0]);

        return 1;

    }

    /*

      Opening input files (file1 and file2) for reading in text mode.

      The output file (OutFile) is open for writing.

    */

    if(((fpin1=fopen(argv[1],"r"))==NULL) ||

       ((fpin2=fopen(argv[2],"r"))==NULL) ||

       ((fpout=fopen(argv[3],"w"))==NULL))

    {

        printf("Error! Could not open filesn");

        return 1;

    }    

    while(fgets(line, sizeof(line), fpin1)!=NULL) /* Read hash line from the first file1 */

    {

        bCheck=FALSE;

        while(fgets(line1, sizeof(line1), fpin2)!=NULL) /* Read hash line from the second file2 */

        {

            if(!strncmp(line, line1, 38)) /* Compares 38 characters of the line in file1 to those of the file2 */

            {

                bCheck=!bCheck;

                break;

            }

        }

        if(!bCheck) /* Does compared line are the same ? */

        {

            fputs(line,fpout); /* Yes - write them in a file OutFile */

            ncHash++; /* Count identical lines */

        }

        rewind(fpin2); /* Seek to the beginning of the file2 */

        count++;      /* Counting the read lines in file1 */

    }

    printf("nDone...n");



    fclose( fpin1);

    fclose( fpin2);

    fclose( fpout);

    return 0;

}   

The OutFile from the program is:

cf03189f5b05eb1a9658f80d7a0e9f02:_#.g}

5701aa8e2aa7e1cfd16ca4076bd1732a:@AQ1z

7a1e6090568e076c55df9dc7abf356c6:9rC@p

04046da33706518d9b15a38bcddb448e:!DFPk

• I want to know how to write it in awk?

awk -F: 'NR==FNR{a[$1,$2];next}!(($1,$2) in a)' excludes.txt all.txt

Note the reversal of file argument order.

An explanation:
-F: - use : as a field separator
NR==FNR - The first file (current row number = total row number)
a[$1,$2] - Touch the array at the first two fields
next - move on to the next line, so we don't have to check the alternate NR != FNR condition
!(($1,$2) in a) - check whether the combination was seen. If not, print the line (default action)

Code for GNU sed:

sed -r 'sµ(.*):.*$µ\§1§dµ' file2 |sed -f - file1

Because of the many 'ugly' characters the code is for info only, do not use in production.

I suppose you mean awk. It's possible to do it, but it will eat 2 times the files sizes of memory:

cat file1 file2 |

awk '{ s = substr($0, 1, 38); str[NR] = s; ex[s]++; } 

END {

       for (i = 1; i <= NR; i++) {

          s = str[i];

          if (ex[s] == 1)

             print s;

       }

}'

Output:

cf03189f5b05eb1a9658f80d7a0e9f02:_#.g}

5701aa8e2aa7e1cfd16ca4076bd1732a:@AQ1z

7a1e6090568e076c55df9dc7abf356c6:9rC@p

04046da33706518d9b15a38bcddb448e:!DFPk

In perl

perl -F: -ane'BEGIN{$f=$ARGV[0]}print if$ARGV ne$f&&!$h{$F[0]};$h{$F[0]}=1' file2 file1