cat or grep a html file to find specific text

Question

I am trying to use bash to parse and HTML file using grep.

The HTML won't change so I should be able to find the text easy enough.

The HTML will be like this, and I just want the number which will change each time the file changes:

<div class="total">
          900 files inspected,
          28301 offenses detected:
        </div>


grep -E '^<div class="total">.</div>' my_file.html

Ideally I just want to pull the number of offenses so in the example above it would be 28301. I would like to assign it to a variable also.

Am I close?

A more *reliable* approach would be something like `tidy -asxml — Charles Duffy, Sep 02 '18 at 22:51
You might also be able to do the entire job without regular expressions or any external tools at all. `x="$( — ghoti, Sep 02 '18 at 23:02
ghoti could you break this down as to whats happening in your example? — user3437721, Sep 03 '18 at 13:14
The syntax used above is described in detail in http://wiki.bash-hackers.org/syntax/pe, and also [covered in BashFAQ #100](http://mywiki.wooledge.org/BashFAQ/100#Removing_part_of_a_string). — Charles Duffy, Sep 03 '18 at 17:33

score 1 · Accepted Answer · answered Sep 02 '18 at 22:08

1

you can do a simple

a=$(grep -oP '(\d+)(?=\soffenses\sdetected)' abc);echo $a

will give:

-o only gives the matching part of the line

-P uses perl regular expression in regex

abc is the name of the file

(\d+)(?=\soffenses\sdetected) in this reges we are just using positive lookahead to capture the require digits that are followed by a particular word

answered Sep 02 '18 at 22:08

Inder

3,711
9
27
42

1

Thanks, makes sense I appreciate the explanation – user3437721 Sep 02 '18 at 22:26
2

Note that this depends on GNU `grep`, and specifically, a GNU `grep` compiled with the optional library libpcre. – Charles Duffy Sep 02 '18 at 22:53
2

To build on the GNU grep note from Charles, I'll point out that `grep` of any type is not actually *part of bash*, and varies from system to system. BSD systems may have an optional `pcregrep` binary installed with a `pcre` package the behaves similarly to `grep -P`, and macOS can get the same tool using Macports or Brew. – ghoti Sep 02 '18 at 22:58
@CharlesDuffy absolutely valid point – Inder Sep 02 '18 at 23:12

score 0 · Answer 2 · answered Sep 02 '18 at 22:27

If you have GNU grep and GNU sed, you can do:

$ cat file | xargs | grep -Po '<div class=total>\K(.*?)</div>' | sed -E 's/<\/div>//; s/, /\n/'
 900 files inspected
28301 offenses detected:

If you have ruby available:

$ ruby -e 'puts readlines.join[/(?<=<div class="total">).+(?=<\/div>)/m].gsub(/^[ \t]+/m,"")' file 
900 files inspected,
28301 offenses detected:

cat or grep a html file to find specific text

2 Answers2