Trying to find a match of two strings in a file using awk

user_7193162 0
Viewed 61

Hello I am trying to find a pattern match on some HTML files using AWK but i dont seem to have any luck with it

So for my pattern to match it should have the following

<tr>
                    <td>Failures</td>
                    <td>0</td>
                </tr>
                <tr>
                    <td>Warnings</td>
                    <td>4</td>
                </tr>
                <tr>
                    <td>Errors</td>
                    <td>0</td>
                </tr>
                <tr>
                    <td>Not Applicable</td>
                    <td>53</td>
                </tr>
                <tr>
                    <td>Manual Checks</td>
                    <td>9</td>
                </tr>

Failures and Manual Checks should be zero. So in the above file failures is 0 and manual check is 9. So i need to match only when failure is 0 and manual check is 0.

SO i tried with and without escaping the new line but awk is not returning any results.

find . -name "*.html" -exec awk '/td\>Failures\<\/td\>\\n.*\<td\>0/ {print FILENAME}' '{}' \;

I have also tried other combinations like below but cant seem to figure out why awk is not going to the next line.

find . -name "*.html" -exec awk '/td\>Failures\<\/td\>\\n\[\^\\\<\]\+\<td\>0/ {print FILENAME}' '{}' \;

Can anyone please have a look and tell me what i am missing?

awk html-parsing
3 Answers

If your html files are well-formed xml, then xmlstarlet will work:

find . -name '*.html' \
       -exec xmlstarlet sel -t \
               --if '//tr[td[1] = "Failures" and td[2] = "0"]' \
               --if '//tr[td[1] = "Manual Checks" and td[2] = "0"]' \
               --inp-name --nl \
             '{}' \;
  • if there's a row where the first cell is Failures and the second cell is 0,
  • and if there's a row where the first cell is Manual Checks and the second cell is 0,
  • then print the input filename and a newline.
user_75520

A more reliable solution is going to be based on a tool designed to parse html; having said that ...

One awk idea using a couple custom regex patterns:

$ cat regex.awk
BEGIN { RS="^$"                                                 # whole file treated as a single line of input
        regex1="<td>Manual Checks</td>[[:space:]]+<td>0</td>"
        regex2="<td>Failures</td>[[:space:]]+<td>0</td>"
      }
$0 ~ regex1 && $0 ~ regex2 {print FILENAME}

NOTE: placing the code in a file (regex.awk) will make the follow-on find/awk quite a bit cleaner

Sample input:

$ cat f1.html
... snip ...
                    <td>Failures</td>
                    <td>0</td>                         # match
... snip ...
                    <td>Manual Checks</td>
                    <td>9</td>                         # not a match
... snip ...

$ cat f2.html
... snip ...
                    <td>Failures</td>
                    <td>0</td>                         # match
... snip ...
                    <td>Manual Checks</td>
                    <td>0</td>                         # match
... snip ...

NOTE: comments added for clarification; comments to not exist in the actual files

Adding this to a find call:

$ find . -name "f?.html" -exec awk -f regex.awk '{}' \;                                                              
./f2.html
user_73661000

Using any awk in any shell on every Unix box:

$ cat tst.awk
gsub("^[[:space:]]*<td>|</td>[[:space:]]*$","") {
    if ( ++cnt % 2 ) {
        tag = $0
    }
    else {
        f[tag] = $0+0
    }
}
END {
    if ( (f["Failures"] == 0) && (f["Manual Checks"] == 0) ) {
        print FILENAME
    }
}


$ awk -f tst.awk file

The above creates an array f[] that maps the tags (names) of the cells to their values so then in the END section you can do whatever test you like on whatever combination of them you like.

user_17450010
Related