grepping several values from a html-page?

[Strowi]

hi,

i am trying to get several values off a website to some bash-script...
this is the part from the <table> where i want to get some values.
The first line is static, followed by 1 or more rows where i want to get the content of the table-data fields...
Anyone got a clue as what the best approach would be?

Code: Select all

<td width="116" bgcolor="#FFCC00"><b>Status</b></td> 
</tr> 
<tr valign="top"> 
<td width="57" bgcolor="#FFFFFF"><img border="0" height="1" width="57" src="../art/pix_white.gif"></td>
<td width="117" bgcolor="">Info. Abgangs-Auswechslungsstelle&nbsp;</td>
<td width="117" bgcolor="">2010-01-06&nbsp;</td><td width="116" bgcolor="">allgemeiner Status.</td> 
</tr> 
</table> 

[Naib]

dont use grep

[cach0rr0]

personally I'd do the whole bit in perl if possible

maybe you could run a few lines of sed over it? still think perl would handle it more gracefully, since you could easily load the whole page into a scalar, remove all but TD tags, then remove all but what's between remaining tags

if that makes sense

[durian]

maybe you could run a few lines of sed over it? still think perl would handle it more gracefully, since you could easily load the whole page into a scalar, remove all but TD tags, then remove all but what's between remaining tags

Doesn't perl even have decent HTML parsers? That would make it even better, a hardcoded regex based on the contents could easily fail if they change a single byte in the file...

-peter

[Naib]

maybe you could run a few lines of sed over it? still think perl would handle it more gracefully, since you could easily load the whole page into a scalar, remove all but TD tags, then remove all but what's between remaining tags

Doesn't perl even have decent HTML parsers? That would make it even better, a hardcoded regex based on the contents could easily fail if they change a single byte in the file...

-peter

exactly.
don't regex over a markup language

[cach0rr0]

maybe you could run a few lines of sed over it? still think perl would handle it more gracefully, since you could easily load the whole page into a scalar, remove all but TD tags, then remove all but what's between remaining tags

Doesn't perl even have decent HTML parsers? That would make it even better, a hardcoded regex based on the contents could easily fail if they change a single byte in the file...

-peter

they do, I'm just a bit hesitant to proffer heaps of info in case this is, say, for a school project or some such
dunno what his requirements are

[Strowi]

whoa lots of replies for this short period..

No, it's not a school project, i need to grep some data from one of our webservers at the university. I am not that familar with perl, guess i will have to dig into that...
Already tried putting it through tiny and awk/sed from the resulting xml, i can even identify the td classes as tiny enumerates them (cXX), but how would i check if there are more than 1 entries ?

greetings,
strowi

[mokona]

Try "lynx -dump".
Then read each column with awk.

[Mike Hunt]

You could try this:

Code: Select all

sed -n '/Status/,/\/table/p' filename.html

[pjp]

Moved from Unsupported Software to Portage & Programming.

[Naib]

Try "lynx -dump".
Then read each column with awk.

For such things I prefer curl or w3m
Then depending on how complex the output is either basic console commands (cut...) or regex commands (grep,sed).. or another lang (py,perl..)

[marens]

I'd say xpath is the way to go.

http://www.w3schools.com/XPath/xpath_syntax.asp

Combine with your favourite programming language.

[Mike Hunt]

This works better, it will only match until the first occurence of table and exit:

Code: Select all

sed -n '/Status/,/table\?/p' file