I want to write a sed (or awk) command to replace a string from one file with the entire contents of another file

Question

I want to write a sed (or awk) command to replace a string from one file with the entire contents of another file. Note that second file from which I want to get the content has more than one line.

$ cat file.txt 
TEXT1 
TEST2 
TEST3 

and

$ cat other_file.txt 
there are multiple lines1 
there are multiple lines2 
there are multiple lines3 

And I want the output to be:

$ cat file3.txt
there are TEXT1 
TEST2 
TEST3 lines1

there are TEXT1 
TEST2 
TEST3 lines2

there are TEXT1 
TEST2 
TEST3 lines3

I tried this:

sed -i -e '/PLACEHOLDER/ r file' -e s/PLACEHOLDER// otherFile

But it didn't give me the right output.

replace "multiple" key word with all the content from file on each line hope this is clear

Answer 1

Using the m4 macro processor, defining a macro called PLACEHOLDER which includes the other file:

$ cat file
foo PLACEHOLDER baz

$ cat otherfile
These are the
multi-line contents
of the other file.

$ m4 -D PLACEHOLDER='include(otherfile)' file
foo These are the
multi-line contents
of the other file.
 baz

Note that this includes the text from otherfile as-is, without stripping the final newline character.

Answer 2

Using any awk in any shell on every Unix box and given 1 possible interpretation of your requirements (that you want to do full-word string matching on space-separated "words", the replacement text might contain backreference metachars or other typically problematic chars, don't care about retaining white space, and that the target word could appear many times on a line):

$ cat tst.awk
BEGIN {
    ORS = RS RS
    old = "multiple"
}
NR==FNR {
    new = (NR>1 ? new RS : "") $0
    next
}
{
    for ( i=1; i<=NF; i++ ) {
        if ( $i == old ) {
            $i = new
        }
    }
    print
}

$ awk -f tst.awk file.txt other_file.txt
there are TEXT1
TEST2
TEST3 lines1

there are TEXT1
TEST2
TEST3 lines2

there are TEXT1
TEST2
TEST3 lines3

Answer 3

Using Raku (formerly known as Perl_6)

~$ raku -e 'my @a = dir(test => "Matsuo_Bashō.txt").IO.lines.join("\n"); for lines[0..9] { put S/e/@a[]/ };'  alphabet.txt
a
b
c
d
No one travels
Along this way but I,
This autumn evening.
f
g
h
i
j

Original_File being substituted == alphabet.txt

~$ raku -e 'for lines[0..9] { .put };' alphabet.txt
a
b
c
d
e
f
g
h
i
j

Second File to be inserted == Matsuo_Bashō.txt

~$ raku -e 'put dir(test => "Matsuo_Bashō.txt").IO.lines.join("\n");'
No one travels
Along this way but I,
This autumn evening.

Briefly, take the file you want to "insert" and read it in using Raku's dir() routine, storing it in the @a array. Then the second 'target' file (here an alphabetic sequence) is read-in off the command line. Raku will output the resultant string when substituted with the S/// "big-S" substitution operator. Note: the code above will only replace the first occurrence. To make this a global replacement (replacing all occurences of the recognition sequence e, add the :global adverb: S:global/// or more simply, S:g///.

There's no need to backslash @a in the replacement, simply add indexing square-brackets at the end: @a[] (curlies work also, see below).

If you don't want to substitute but instead want to insert a second file after a recognition sequence in the first, change the substitution operator to use a lookbehind assertion, and add a newline before the replacement:

{ put S/ <?after e> /\n@a[]/ }

#OR

{ put S/ <?after e> /\n{@a}/ }

Finally note that files can be slurped-in all at once. Using Raku's dir() routine that's:

• dir(test => "Matsuo_Bashō.txt").IO.slurp;

...however an extra newline gets added to the output. You can use slurp and trim-trailing to replicate the answer above (below, a shorthand way of quoting the dir() filename string is to enclose it in angle brackets):

~$ raku -e 'my @a = dir(:test<Matsuo_Bashō.txt>).IO.slurp; for lines[0..9] { put S/e/{@a[].trim-trailing}/ };'  alphabet.txt

https://raku.org