Sed / Grep-过滤出内容

Question

Given that I have something like this:

int val; struct node *mhm;
int val; struct node *next;

I want to extract content on each line, using sed, awk, or grep so that they read:

init->val = val; init->mhm = mhm;
init->val = val; init->next = next;

grep -o outputs each individual result on a new line, but I need to preserve line structure (each group of variables refer to a separate thing).

I use a gnu sed command presently, but I only want to output the replaced string:

sed -re 's/([A-Za-z0-9_]*);/init->\1 = \1;/g'

This above sed command outputs:

int init->val = val; struct node *init->mhm = mhm;
int init->val = val; struct node *init->next = next;

Even when using non-greedy modifiers, I can't get a full line regex statement to successfully select each variable name in each line.

Answer 1

这适用于您的样本集（对解决方案进行了微小更改）：

sed -re 's/[^;]* [*]?([A-Za-z0-9_]+) *;/init->\1 = \1; /g'

Answer 2

There isin't a variable name in each line unless you parse the language.

That being said, as just random text, this:
s/([^\\S\\n]*)[^;\\n]+?(\\w+)\\s*;/\\1init->\\2 = \\2;/g
should do it.

Answer 3

尝试这个

awk '{for(i=1;i<=NF;i++) if($i~/;/){ gsub(/\*/,"",$i);a=gensub(/(\w*);/, "init->\\1 = \\1; ","g",$i);printf a;}print ""}' yourFile