CODEWITHSUNDEEP
awksed
Bastien974
Bastien974

Reputation: 321

Pipe output of awk to sed in single command

I have a simple file like :

a 123
b 234

In a second file, I want to replace a string by the value corresponding to a -> "123" in a single command.

Is there a way to pipe the result of grep "a" file1 | awk {print $2} to sed s/string/my_output/g file2

Upvotes: 2

Views: 6428

Answers (4)

Jotne
Jotne

Reputation: 41446

Here is an awk

cat file1
a 123
b 234


cat file2
cat
bad hat


awk -v FS="" 'FNR==NR {split($0,b," ");a[b[1]]=b[2];next} {for (i=1;i<=NF;i++) $i=a[$i]?a[$i]:$i}1' file1 file2
c 123 t
234 123 d   h 123 t

How does it work.

awk -v FS="" '                      # By setting Field Separator to nothing, it works on one and one letter
FNR==NR {                           # This is true only for the first file (file1)
    split($0,arr-b," ")             # Split the data in first file (arr-b[1]=first filed, eks a, arr-b[2]=second field, eks 123
    arr-a[arr-b[1]]=arr-b[2]        # Store this to an array using field 1 as index. arr-a[a]=123 arr-a[b]=234 etc
    next}                           # Skip to next record in file1
    {                               # Run this for file2
    for (i=1;i<=NF;i++)             # Loop trough on by on letter "c" "a" "t"
        $i=arr-a[$i]?arr-a[$i]:$i}  # Test the letter one by one if its found in array "arr-a", if so use data from array, if not keep value
1                                   # Since this always will be true, do default action "print $0" (print every line)
' file1 file2                       # Read file1 and file2

Upvotes: 1

johntellsall
johntellsall

Reputation: 15160

here's a pure-Awk solution. Put the dictionary in 'adict.dat' then pass your input file to the Awk command

source

BEGIN { 
    while(( getline line < "adict.dat" ) > 0 ) {
        $0 = line;
        myvars[$1] = $2;
    }
}

{
    for (varname in myvars) {
        gsub(varname, myvars[varname]);
    }
    print;
}

Example run

/bin/echo 'a x b y' | gawk -f arepl.awk

output

123 x 234 y

Upvotes: 0

Zac Lozano
Zac Lozano

Reputation: 818

You could do the following:

grep a file1| awk '{print $2}' > tmp.txt

This will put '123' in a temporary file. Then:

sed -f script.sed file2

Where the contents of script.sed is

/string to replace/ {
    r tmp.txt
    d
}

This script replaces "string to replace" with the contents of tmp.txt.

Upvotes: 0

shellter
shellter

Reputation: 37258

If I interpret your question correctly, try

sed 's@^\(.*\) \(.*\)$@s/\1/\2/g@' SimpleFile > substitutions.file
sed -f substitutions.file 2ndFile > 2ndFile.tmp && mv 2ndFile.tmp 2ndFile

The first step will create a sed substitions file like

s/a/123/g
s/b/234/g

which when used as in line2 above, will substitute each a with 123, etc.

IHTH

Upvotes: 0

Related Questions