CODEWITHSUNDEEP
unixawkgrep
user2966591
user2966591

Reputation: 93

Delete certain rows in text files using UNIX

I have a bunch of text files that needs cleaning. Im using UNIX bash, so AWK or grep is good.

The text files looking something like this: 

1766 1789  
1764 1790  
1762 1849  
0  
1357 1817  
1366 1857  
0  
360 42  
352 95  
0  
293 142  
302 181  
delete-this  
0  
302 181   
0

What I want is to delete all rows with "0", "delete-this", only one row with two columns or three rows with two columns.

The result should look like this: 

1766 1789    
1762 1849   
1357 1817  
1366 1857    
360 42  
352 95    
293 142  
302 181

Thanks a lot!

More info: The sum of row 1 column 2 and row 2 column 2 should be >1, if not, row 2 must be deleted.

Upvotes: 0

Views: 423

Answers (1)

Jotne
Jotne

Reputation: 41446

This was a hard nut, or difficult to understand, but here we go again:

awk '/[0-9]+ [0-9]+/ {a[++t]=$0;b[t]=$2;next} {if (t>=2) for (i=1;i<=t;i++) {if (b[i]-c!=1) print a[i];c=b[i]};t=0}'
1766 1789
1762 1849
1357 1817
1366 1857
360 42
352 95
293 142
302 181

How does it work:

awk '
    /[0-9]+ [0-9]+/ {               # if line does have 2 column of number, then 
        a[++t]=$0                   # add line to array "a" and increment variable "t"
        b[t]=$2                     # add column 2 to array "b"
        next                        # go to next line
        }

        {
        if (t>=2)                   # is there more two or more lines with numbers connrected, then
            for (i=1;i<=t;i++) {    # loop trough array "a" with all numbers
                if (b[i]-c!=1)      # test if the difference between this number in column 2 is more than 1 compare to previous line
                    print a[i]      # then print array "a"
                    c=b[i]          # store array "b" information in variable "b"
                }
            ;t=0                    # clear counter "t"
        }' file

Upvotes: 2

Related Questions