Cut out a date/time stamp by using SED and scripting

studdard · 03-01-2009, 03:18 AM

Hello

I hope somebody will be able to give me a glue how to cut out a date time stamp out of a modifyed thunderbird adress file.

Background: I am using a small add-on to backup my thunderbird adress book to an external server. I want to use those files to sync my PDA with finchsynch afterwards.

At this point I believe the tool SED will do this job - just do not know how this is possible. A simple search and replace can not do this as the Date/Time stamp is changing all the the time.

(At a later stage I also want to do the opposite way and ..... but first I need this going )

**** file where the time stamp need to be cut out ****
//  Sun Feb 01 2009 21:14:27 GMT+0100 (CET)
< <(a=c)> // (f=iso-8859-1)

**** file how it should look like ****
// 
< <(a=c)> // (f=iso-8859-1)

Thanks for any help

studdard

H_TeXMeX_H · 03-01-2009, 04:08 AM

Probably this will work:

Code:

sed 's|-->.*|-->|g'

sycamorex · 03-01-2009, 04:09 AM

You can try the following:

Quote:

sed -e 's/-->.*/-->/' file

JulianTosh · 03-01-2009, 04:15 AM

Code:

$ echo '// <!-- <mdb:mork:z v="1.4"/> --> Sun Feb 01 2009 21:14:27 GMT+0100 (CET)' | sed 's/^\/\/.*mdb:mork.*--> \(...\) \(...\) \(..\) \(....\) \(..:..:..\).*/\4 \2 \3 \1 \5/'
2009 Feb 01 Sun 21:14:27

theNbomr · 03-01-2009, 11:58 AM

To help learn how to understand the regex needed to perform this kind of operation, start by trying to express in words, the most generalized translation that will work. In your case, I would say something like "remove all characters between the string '-->' and the end of the string." Having done this much, the regular expression that performs this operation almost writes itself. This applies, of course, to the semantic aspects of the regex. The syntax, well, you just have to learn and memorize that part.
--- rod.