如何限制awk只搜索某个HTML标签中包含的项目？

Question

我有一个AWK这样的脚本，我将在一个文件上运行：

cat input.txt | awk 'gsub(/[^ ]*(fish|shark|whale)[^ ]*/,"(&)")' >> output.txt 

这增加了括号包含文字“鱼”，“鲨鱼”的所有行，或“鲸鱼”，为如：

The whale asked the shark to swim elsewhere. 
The fish were unhappy. 

通过脚本运行它后，文件就变成了：

The (whale) asked the (shark) to swim elsewhere. 
The (fish) were unhappy. 

的文件标有HTML标签，我需要做的只是更换之间发生<b>和</b>标签。

The whale asked <b>the shark to swim</b> elsewhere. 
<b>The fish were</b> unhappy. 

这将成为：

The whale asked <b> the (shark) to swim </b> elsewhere. 
<b> The (fish) were </b> unhappy. 

• 匹配大胆的标签永远放在不同的行。起始<b>标记总是与结尾</b>标记在同一行上。

我怎样才能限制awk的搜索只搜索和修改<b>和</b>标签之间的文本？

Answer 1

下面是使用awk的技术：

awk '/<b>/{f=1}/<\/b>/{f=0}f{gsub(/fish|shark|whale/,"(&)")}1' RS=' ' ORS=' ' file 
The whale asked <b>the (shark) to swim</b> elsewhere. 
<b>The (fish) were</b> unhappy. 

Answer 2

只要HTML标记是不是比这更糟糕，而<b> ... </b>跨度将不包含任何其他HTML标记，那么就比较容易在Perl：

$ cat data 
The whale asked <b>the shark to swim</b> elsewhere. 
<b>The fish were</b> unhappy. 
The <b> dogfish and the sharkfin soup</b> were unscathed. 
$ perl -pe 's/(<b>[^<]*)\b(fish|shark|whale)\b([^<]*<\/b>)/\1(\2)\3/g' data | so 
The whale asked <b>the (shark) to swim</b> elsewhere. 
<b>The (fish) were</b> unhappy. 
The <b> dogfish and the sharkfin soup</b> were unscathed. 
$ 

我试图适应，要awk（和gawk），并没有成功;比赛部分工作，但替换表达没有。阅读本手册，与Perl不同，您无法在圆括号中标识单独的匹配子表达式。