UNIX for Dummies Questions & Answers

Hi all, I just signed up to the forums, although, I have lurked on here for awhile. Anyways, my issue is I am trying to get awk to spit out something I can use without having to spend hours in excel hell haha. So, I used sed to replace the spaces with semicolons and redirected that to a file.


This is what I intend to do:

Code:

User1 id=11111 gecos=John Doe groups=group1,group2 account_locked=false

But, because there is a space between the name in the gecos column, it messes everything up. So I need some way to combine the columns starting with gecos=, ignore the space, and end with a character a-z/A-Z and not get to the groups= column. I cannot for the life of me think how to translate this to a command. So, is there any easy way to go about this? It doesn't have to be awk, I just need something that will work. Thanks in advance.

How about sed:

Code:

$ sed -e :a -e 's/\(.*\) \([a-z_]\{1,\}=\)/\1;\2/;ta' file
User1;id=11111;gecos=John Doe;groups=group1, group2;account_locked=false

Or just the values with awk:

Code:

$ awk -F' *[a-z_]+=' '{$1=$1}1' OFS=';'  file
User1;11111;John Doe;group1, group2;false

Awesome, thank you so much Scrutinizer, that works perfectly! I will now know next time how to do this.

Code:

$ echo "User1 id=11111 gecos=John Doe groups=group1,group2 account_locked=false" | sed "s/ \([^= ]\{1,\}=\)/;\1/g"
User1;id=11111;gecos=John Doe;groups=group1,group2;account_locked=false

@anbu, you are right it could also be done from left to right instead of right to left (in fact the = "anchor" provides mini-right to lefts) The [^=] might not even be required and an asterisk gets rid of more than one space:

Code:

sed 's/ *\([^ ]\{1,\}=\)/;\1/g'

Just another one.

Code:

sed 's/[^ ]*=/;&/g;s/ ;/;/g' yourfile

or to deal with the "more than one space" issue :

Code:

sed 's/[^ ]*=/;&/g;s/ *;/;/g' yourfile

All excellent information, thanks all. This has saved me so much time.