CODEWITHSUNDEEP
regexstringbash
xbmono
xbmono

Reputation: 2316

How to extract the first and last two parts from a string separated by underscore in bash?

So we have a file named like this:

some-app-name_CT-111-some-title-with_underscore-in-it_c37a9a5fc272a5c94009a61ce8dff79900ab9102_2017-07-24-03-22-19.tar.bz2

As you can see there are 4 parts, a app name (dasherized), a title (that may contain underscore in it), a hash code and finally a timestamp (dasherized)

They are separated by underscore, the problem is that the title may have underscore in it. So how can we get the first part first and then the last two parts (separated by underscore) and then the remain is the title?

Any help is appreciated.

Final parts should be like:

Upvotes: 1

Views: 141

Answers (2)

anubhava
anubhava

Reputation: 784888

Using bash regex you can do this:

s='some-app-name_CT-111-some-title-with_underscore-in-it_c37a9a5fc272a5c94009a61ce8dff79900ab9102_2017-07-24-03-22-19.tar.bz2'

re='^([^_]+)_([a-zA-Z0-9_-]+)_([a-fA-F0-9]+)_([0-9-]+)\.'

[[ $s =~ $re ]] && printf "AppName: %s\nTitle: %s\nID: %s\nTimestamp: %s\n" \
        "${BASH_REMATCH[1]}" "${BASH_REMATCH[2]}" "${BASH_REMATCH[3]}" "${BASH_REMATCH[4]}"

AppName: some-app-name
Title: CT-111-some-title-with_underscore-in-it
ID: c37a9a5fc272a5c94009a61ce8dff79900ab9102
Timestamp: 2017-07-24-03-22-19

Upvotes: 2

Guru
Guru

Reputation: 16974

One way:

appname=$(echo $x | awk -F_ '{print $1}')
hcode=$(echo $x | awk -F_ '{print $(NF-1)}')
timestamp=$(echo $x | awk -F_ '{print $NF}' | grep -oE '[0-9-]{2,}')
title=$(echo $x | sed "s/.*${appname}_\(.*\)_$hcode.*/\1/")

where x is the variable containing the filename

hashcode is retrieved by fetching the second last column with _ as delimiter. timestamp is retrieved from the last column and extracting only numbers and - out of it. Title is retrieved by fetching characters between the appname and the hashcode.

Upvotes: 1

Related Questions