在Awk中使用RegExp

Question

I have CSV file:

<iframe src="https://localhost/get/44bc40f3bc04f65b7a35"></iframe>|name_1
<iframe src="https://localhost/get/5db0d477d707121934ff"></iframe>|name_2
<iframe src="https://localhost/get/6c95bd2b32ed45989c61"></iframe>|name_3
<iframe src="https://localhost/get/0a9c4655800e8a7b9ea2"></iframe>|name_4
<iframe src="https://localhost/get/754953b57a32e2841bda"></iframe>|name_5

and want use RegExp and Awk (or Gawk) to make this CSV file like this:

44bc40f3bc04f65b7a35|name_1
5db0d477d707121934ff|name_2
6c95bd2b32ed45989c61|name_3
0a9c4655800e8a7b9ea2|name_4
754953b57a32e2841bda|name_5

I have worked RegExp in Grep

$ grep -Po "[A-Za-z]*+\d++\w++" example.txt 
44bc40f3bc04f65b7a35
5db0d477d707121934ff
6c95bd2b32ed45989c61
0a9c4655800e8a7b9ea2
754953b57a32e2841bda

but this RegExp not work in Awk. I think i'm not correct use regexp in Awk or this type of RegExp not worked in Awk.

$ awk -F "|" 'match($1, /[A-Za-z]*+\d++\w++/, a) {print a[0]"|"$2}' example.txt 
db0d477d707121934ff|name_2
bd2b32ed45989c61|name_3
bda|name_5

Just Awk work fine:

$ awk -F "|" '{print $1"|"$2}' example.txt 
<iframe src="https://localhost/get/44bc40f3bc04f65b7a35"></iframe>|name_1
<iframe src="https://localhost/get/5db0d477d707121934ff"></iframe>|name_2
<iframe src="https://localhost/get/6c95bd2b32ed45989c61"></iframe>|name_3
<iframe src="https://localhost/get/0a9c4655800e8a7b9ea2"></iframe>|name_4
<iframe src="https://localhost/get/754953b57a32e2841bda"></iframe>|name_5

Answer 1

Try:

$ awk -F'<iframe src="https://localhost/get/|"></iframe>' '{print $2 $3}' file
44bc40f3bc04f65b7a35|name_1
5db0d477d707121934ff|name_2
6c95bd2b32ed45989c61|name_3
0a9c4655800e8a7b9ea2|name_4
754953b57a32e2841bda|name_5

This works by setting the input field separator to be either iframe src="https://localhost/get/ or "></iframe> in which case the output you want is the second field followed by the third field.

Alternative Method Using Match

$ awk -F "|" 'match($1, /[[:xdigit:]]{20}/, a) {print a[0]"|"$2}' file
44bc40f3bc04f65b7a35|name_1
5db0d477d707121934ff|name_2
6c95bd2b32ed45989c61|name_3
0a9c4655800e8a7b9ea2|name_4
754953b57a32e2841bda|name_5

Note that awk supports POSIX regular expressions. That means that it recognizes character classes like [[:digit:]] or [[:alnum:]] but not necessarily \\d or \\w . As a GNU-specific extension, gawk supports \\w (but not \\d ). For portability, stick to the POSIX classes as shown in man 7 regex .

Yet another method

Your regex [A-Za-z]*+\\d++\\w++ can be translated into awk as follows:

$ awk -F "|" 'match($1, /[[:alpha:]]*[[:digit:]]+[[:alnum:]]+/, a) {print a[0]"|"$2}' file
44bc40f3bc04f65b7a35|name_1
5db0d477d707121934ff|name_2
6c95bd2b32ed45989c61|name_3
0a9c4655800e8a7b9ea2|name_4
754953b57a32e2841bda|name_5

Note that this method requires that the 20-character hex string contains at least one digit.

Answer 2

1. The difference between awk and grep invocations in your example is -P option in grep, which stands for "Use Perl regexp". If you replace it with -E, it will work just like your awk run. Awk does not support Perl extension.

2. Your regexp is better be fixed, I don't think you need these extra + signs, to begin with. If I can assume that you need all letters or digits after get/ then I'd rather write:

   awk -F "|" 'match($1, /get/([A-Za-z0-9]+)/, a) {print a[1]"|"$2}' example.txt

Here we use [A-Za-z0-9]+ match any number of small or upper letters or digits that come after /get, use a[1] to print a matched group inside the parentheses instead of the whole matching pattern a[0] which includes get/

Answer 3

awk '{gsub(/<.*get\//,"")gsub(/".*e>/,"")}1'  file

44bc40f3bc04f65b7a35|name_1
5db0d477d707121934ff|name_2
6c95bd2b32ed45989c61|name_3
0a9c4655800e8a7b9ea2|name_4
754953b57a32e2841bda|name_5

Answer 4

Here is another solution:

awk -F"[/\">|]" 'BEGIN{ OFS = "|" }{ print $6, $11 }' yourfile

With the -F option in the beginning the Field Separator can be /, ", > and |. After that is done you can just print the fields $6 and $11 which contain your desired output together with the output field separator.

Output:

44bc40f3bc04f65b7a35|name_1
5db0d477d707121934ff|name_2
6c95bd2b32ed45989c61|name_3
0a9c4655800e8a7b9ea2|name_4
754953b57a32e2841bda|name_5