How to delete all instances of "<a href=(maybe a no/ letter/ sign)</a>" from a file.

Hello folks,
Again have hundreds of large file, from where i want to delete all internal and external link with their texts. Any idea about how can i do so with SED or ant other utility in commandline or script.

Basically i have to delete

Code:

---

<a href=(maybe a no/ letter/ sign)</a>

---

. The bracketed portion is highly changeable so looking for a sophisticated wildcard actually.

Thanks for your time.

Code:

---

sed -ie 's|<[a-zA-Z] href.*\/[a-zA-Z]>||' fileOrFiles

---

I think i have made a little mistake. Sorry. Actually the string patterns are

Code:

---

<a href="#[1-9]|[1-4][0-9]|50">See [1-9]|[1-4][0-9]|50</a>

---

and

Code:

---

<a href="#[1-9]|[1-4][0-9]|50">See [1-9]|[1-4][0-9]|50</a> and <a href="#[1-9]|[1-4][0-9]|50"> [1-9]|[1-4][0-9]|50</a>

---

sed have to delete both the 1st one and the 2nd one.

Thanks guys for your time n help.

Quote:

---

Originally Posted by HiImTye

Code:

---

sed -ie 's|<[a-zA-Z] href.*\/[a-zA-Z]>||' fileOrFiles

---

---

will grab

Code:

---

<a href=blah blah junk></a>
<Z hrefblahblahblah/q>

---

etc

I guess that "e" in the command is an trouble besides your code does remove some <a href= ...../> but deleting whole lines after the link anchor and it is also not working properly to remove all instances of <a href=**>.

I didn't undrrstand what your last comment was. give some specific examples of what yoi need that isn't being done, or you need done diffrrently, and I can give you some examples of how to achieve them.

To catch and strip only <a></a> tags i'd suggest:

Code:

---

sed -e 's|</\?a[^<>]*>||g' < infile > outfile

---

Should work in most situations.

Ya it is working nicely. Can you please describe the code a lil bit, so i can use them in informed way.

Thanks for your time and help.

Well, the s command tells sed we're doing pattern replacement (substitution). | characters separate the pattern to math, substitution string (which is empty in this case) and sed commands. Finally the g command tells sed to do more that one substitution per line if necessary.

</\?a[^<>]*> is the patthern we're looking for. Outer < and > are literals. /\? matches slash or nothing. a is literal. [^<>]* matches zero or more occurrences of anything that isn't < or > (regular expressions are greedy so without this part sed would remove a lot more than we wanted).

If you want to learn sed, start here (nice tutorial with lots of examples):
http://www.grymoire.com/Unix/Sed.html