如果在文件 2 中找不到字符串，则用空格替换文件 1 中的字符串

Question

This question: https://unix.stackexchange.com/questions/20322/replace-string-with-contents-of-a-file-using-sed replaces a fixed string in file1 with the contents of file 2.

I want to do this the other way around plus an inversion.

If I have file1:

A:B
B:B
C:
D:
E:A

and file2:

D
E
:

then I want to be left with

:
:
:
D:
E:

If anyone has any pointers that would be great. Bonus points if this can be done on a specific column of a file1 while preserving the rest of the file1.

ie If I have three columns:

A:B    A:B    A:B
B:B    B:B    B:B
C:     C:     C:
D:     D:     D:
E:A    E:A    E:A

I would end up with (target column 2)

A:B    :      A:B
B:B    :      B:B
C:     :      C:
D:     D:     D:
E:A    E:     E:A

Answer 1

tr makes this trivial:

$ tr -cd "$(cat file2)" < file1         
:
:
:
D:
E:

Answer 2

$ cat tst.awk
BEGIN { FS=OFS="\t" }
NR == FNR {
    goodChars[$1]
    next
}
{
    goodStr = ""
    for (i=1; i<=length($2); i++) {
        char = substr($2,i,1)
        if (char in goodChars) {
            goodStr = goodStr char
        }
    }
    $2 = goodStr
    print
}

$ awk -f tst.awk file2 file1
A:B     :       A:B
B:B     :       B:B
C:      :       C:
D:      D:      D:
E:A     E:      E:A

The above assumes your input file is tab-separated as it looks like it is, otherwise just get rid of the BEGIN section.

Answer 3

This might work for you (GNU sed):

sed -z 's/\n//g;s/.*/s#[^&]##g/' file2 | sed -f - file1

Convert file2 into a sed script and run it against file1. This concatenates each character in file2 and places them in a negative character class inside a sed substitution command which runs globally ie the command removes all occurrences of any character in file2 from file1.

To cater for the second problem, add newlines to the negative character class, isolate the second column, make a copy, apply the same code and using pattern matching replace the second column with the amended value:

sed -z 's/\n//g;s/.*/s#[^&\\n]##g/' file2 |
sed -Ee 's/\S+/\n&\n/2;h' -f - -e 'H;g;s/\n.*\n(.*)\n.*\n(.*)\n/\2\1/' file3