How to extract data from html table in shell script?

Answer 1

Go with (g)awk, it's capable :-), here is a solution, but please note: it's only working with the exact html table format you had posted.

 awk -F "</*td>|</*tr>" '/<\/*t[rd]>.*[A-Z][A-Z]/ {print $3, $5, $7 }' FILE

Here you can see it in action: https://ideone.com/zGfLe

Some explanation:

1. -F sets the input field separator to a regexp (any of tr's or td's opening or closing tag

2. then works only on lines that matches those tags AND at least two upercasse fields

3. then prints the needed fields.

HTH

Answer 2

A solution based on multi-platform web-scraping CLI xidel and XQuery:

xidel -s --xquery 'for $tr in //tr[position()>1] return join($tr/td, " ")' file

With the sample input, this yields:

SAVE_DOCUMENT OK 0.406 s
GET_DOCUMENT OK 0.332 s
DVK_SEND OK 0.001 s
DVK_RECEIVE OK 0.001 s
GET_USER_INFO OK 0.143 s
NOTIFICATIONS OK 0.001 s
ERROR_LOG OK 0.001 s
SUMMARY_STATUS OK 0.888 s

Explanation:

• XQuery query for $tr in //tr[position()>1] return join($tr/td, " ") processes the tr elements starting with the 2nd one (position()>1, to skip the header row) in a loop, and joins the values of the child td elements ($tr/td) with a single space as the separator.

• -s makes xidel silent (suppresses output of status information).

---

While html2text is convenient for display of the extracted data, providing machine-parseable output is non-trivial, unfortunately:

html2text file | awk -F' *\\|' 'NR>2 {gsub(/^\||.\b/, ""); $1=$1; print}'

The Awk command removes the hidden \b-based (backspace-based) sequences that html2text outputs by default, and parses the lines into fields by |, and then outputs them with a space as the separator (a space is Awk's default output field separator; to change it to a tab, for instance, use -v OFS='\t').

^{Note: Use of -nobs to suppress backspace sequences at the source is not an option, because you then won't be able to distinguish between the hidden-by-default _ instances used for padding and actual _ characters in the data.}

Note: Given that html2text seemingly invariably uses | as the column separator, the above will only work robustly if the are no | instances in the data being extracted.

Answer 3

You may use html2text command and format the columns via column, e.g.:

$ html2text table.html | column -ts'|'

Component                                      Status  Time / Error
SAVE_DOCUMENT                                           OK            0.406 s     
GET_DOCUMENT                                            OK            0.332 s     
DVK_SEND                                                OK            0.001 s     
DVK_RECEIVE                                             OK            0.001 s     
GET_USER_INFO                                           OK            0.143 s     
NOTIFICATIONS                                           OK            0.001 s     
ERROR_LOG                                               OK            0.001 s     
SUMMARY_STATUS                                          OK            0.888 s     

then parse it further from there (e.g. cut, awk, ex).

In case you'd like to sort it first, you can use ex, see the example here or here.

Answer 4

You can use bash xpath (XML::XPath perl module) to accomplish that task very easily:

xpath -e '//tr[position()>1]' test_input1.xml 2> /dev/null | sed -e 's/<\/*tr>//g' -e 's/<td>//g' -e 's/<\/td>/ /g'

Answer 5

You can parse the file using Ex editor (part of Vim) by removing HTML tags, e.g.:

$ ex -s +'%s/<[^>]\+>/ /g' +'v/0/d' +'wq! /dev/stdout' table.html 
  SAVE_DOCUMENT  OK  0.406 s  
  GET_DOCUMENT  OK  0.332 s  
  DVK_SEND  OK  0.001 s  
  DVK_RECEIVE  OK  0.001 s  
  GET_USER_INFO  OK  0.143 s  
  NOTIFICATIONS  OK  0.001 s  
  ERROR_LOG  OK  0.001 s  
  SUMMARY_STATUS  OK  0.888 s 

Here is shorter version by printing the whole file without HTML tags:

$ ex +'%s/<[^>]\+>/ /g|%p' -scq! table.html

Explanation:

• %s/<[^>]\+>/ /g - Substitute all HTML tags into empty space.
• v/0/d - Deletes all lines without 0.
• wq! /dev/stdout - Quits editor and writes the buffer to the standard output.

Answer 6

There are a lot of ways of doing this but here's one:

grep '^<tr><td>' < $FILENAME \
| sed \
    -e 's:<tr>::g'  \
    -e 's:</tr>::g' \
    -e 's:</td>::g' \
    -e 's:<td>: :g' \
| cut -c2-

You could use more sed(1) (-e 's:^ ::') instead of the cut -c2- to remove the leading space but cut(1) doesn't get as much love as it deserves. And the backslashes are just there for formatting, you can remove them to get a one liner or leave them in and make sure that they're immediately followed by a newline.

The basic strategy is to slowly pull the HTML apart piece by piece rather than trying to do it all at once with a single incomprehensible pile of regex syntax.

Parsing HTML with a shell pipeline isn't the best idea ever but you can do it if the HTML is known to come in a very specific format. If there will be variation then you'd be better with with a real HTML parser in Perl, Ruby, Python, or even C.