awk search pattern in a specific field and replace its content - awk

Question

I need to found field of password that is empty, with space or tab, and replace it with x (on /etc/passwd file)
I found this syntax with awk, that show users where second field (using : as delimiter) is or empty, or has space or tab inside:
awk -F":" '($2 == "" || $2 == " " || $2 == "\t") {print $0}' $file
and result is the follow:
user1::53556:100::/home/user1:/bin/bash
user2: :53557:100::/home/user2:/bin/bash
user3: :53558:100::/home/user3:/bin/bash
How I can say to awk to replace this 2nd field (empty or with space or tab) with another character? (for example x)

Answer 1

Could you please try following.
awk 'BEGIN{FS=OFS=":"} {$2=$2=="" || $2~/^[[:space:]]+$/?"X":$2} 1' Input_file
Explanation: Adding explanation of above code.
awk ' ##Starting awk program here.
BEGIN{ ##Starting BEGIN section here which will be executed before Input_file is being read.
FS=OFS=":" ##Setting FS and OFS as colon here for all lines of Input_file.
} ##Closing BEGIN section block here.
{
$2=$2=="" || $2~/^[[:space:]]+$/?"X":$2 ##Checking condition if $2(2nd field) of current line is either NULL or having complete space in it then put its vaklue as X or keep $2 value as same as it is.
}
1 ##mentioning 1 will print edited/non-edited current line.
' Input_file ##Mentioning Input_file name here.
EDIT: As per OP, OP need NOT to touch last line of Input_file so adding following solutio now.
tac Input_file | awk 'BEGIN{FS=OFS=":"} FNR==1{print;next} {$2=$2=="" || $2~/^[[:space:]]+$/?"X":$2} 1' | tac
EDIT2: In case you want to do it kin single awk itself then try following.
awk '
BEGIN{
FS=OFS=":"
}
prev{
num=split(prev,array,":")
array[2]=array[2]=="" || array[2]~/^[[:space:]]+$/?"X":array[2]
for(i=1;i<=num;i++){
val=(val?val OFS array[i]:array[i])
}
print val
val=""
}
{
prev=$0
}
END{
if(prev){
print prev
}
}' Input_file
In case you want to change Input_file itself append > temp_file && mv temp_file Input_file in above code.

Answer 2

$ awk 'BEGIN{FS=OFS=":"} (NF>1) && ($2~/^[[:space:]]*$/){$2="x"} 1' file
user1:x:53556:100::/home/user1:/bin/bash
user2:x:53557:100::/home/user2:/bin/bash
user3:x:53558:100::/home/user3:/bin/bash
To change the original file using GNU awk:
awk -i inplace 'BEGIN{FS=OFS=":"} (NF>1) && ($2~/^[[:space:]]*$/){$2="x"} 1' file
or with any awk:
awk 'BEGIN{FS=OFS=":"} (NF>1) && ($2~/^[[:space:]]*$/){$2="x"} 1' file > tmp && mv tmp file
The test for NF>1 ensures we only operate on lines that already have at least 2 fields and so we don't create a line like :x in the output when there's an empty line in the input file. The rest is hopefully obvious.