Shell Programming and Scripting

I have a file, each line has the date and time twice, once at the start of the line, and again half way along. to neaten things up, and to make it easier to read i'm removing one set.

Wasn't as easy as identify the text and remove, as it'd remove both.

So i added some text at the beginning of the line (word, Remove)

removing the time works using:

Code:

sed -r 's/remove [0-9]{2}:[0-9]{2}:[0-9]{2} //' myfile

Removing the date however doesnt work

Code:

sed -e "s/remove [A-Z][a-z]{3} [0-9]{2}//" myfile

The date is in the format mmm dd and uses title case - ie Jun
I'm assuming my problem is identifying upper and lower case?

help! (please )

You may need to change

Code:

[A-Z][a-z]{3}

to

Code:

[A-Za-z]{3}

because your statement would look for first [A-Z] and then 3 lowercase characters.


If it does not work then try this:

Code:

sed -e 's/remove [A-Z][a-z][a-z] [0-9][0-9]//' myfile

Last option worked.. strangely it didnt work until i switched

Code:

[0-9]{2}

with

Code:

[0-9][0-9]

no idea why it'd have an issue with that element.

Sometimes you have to escape the curly brackets with a backslash to have the quantifier work:

Code:

$ echo AxxA | sed 's/x{2}/_/'
AxxA
$ echo AxxA | sed 's/x\{2\}/_/'
A_A

"Sometimes" is: sed and grep.
They use Unix RE or BRE.
Unlike egrep and awk: they use ERE.
Perl5 uses ERE plus own extensions.