Bash include string between chars in awk print

Question

I'm trying to parse a log file that will have lines like this:

aaa bbb ccc: [DDD] efg oi    
aaa bbb ccc: lll [DDD] efg oo    
aaa bbb ccc: [DDD]

where [DDD] can be at any place in line.

Only one thing will be between [ and ] in any line

Using awk and space as a delimiter, how can I print 1st, 3rd and all data (whole string) between [ and ]?

Expected output: aaa ccc: DDD

What have you tried, and how did it fail? It seems like you should be able to solve this by reading about half of the first page of any Awk introduction. — tripleee, Apr 25 '17 at 13:25
`awk '{print $1,$3, ???}'`, to print data at certain possitions, but I'm asking for help with including data between chars that can be at any place in string. — Tadija Bagarić, Apr 25 '17 at 13:26
@TadijaBagarić: Did you look at similar questions before posting here? — Inian, Apr 25 '17 at 13:33
@Inian I found a usefull answer here: `http://stackoverflow.com/a/21077990/3190306` but need help using it to get what I need — Tadija Bagarić, Apr 25 '17 at 13:34
If you have `aaa bbb [ccc:] DDD` do you want `aaa ccc:` or `aaa ccc: ccc:`? What about `aaa [bbb] ccc: DDD`? It could be in field order of `aaa bbb ccc:` or it could be as you stated order of `aaa ccc: bbb` — dawg, Apr 25 '17 at 14:58
I have no such instances in my file, but your solution will definitely come in handy! — Tadija Bagarić, Apr 25 '17 at 15:18
@TadijaBagarić Don't edit 1 typo in a post if there are many more errors there, you are the reason the edit queue is full (http://stackoverflow.com/questions/44024001/how-do-i-refer-a-control-or-view-in-an-item-of-listview-xamarin-forms#44024001) — EpicKip, May 17 '17 at 12:33

RomanPerekhrest · Accepted Answer · 2017-04-25T14:13:45.097

3

gawk(GNU awk) approach:

Let's say we a file with the following line:

aaa bbb ccc: ddd [fff] ggg hhh

The command:

awk '{match($0,/\[([^]]+)\]/, a); print $1,$3,a[1]}' file

The output:

aaa ccc: fff

match(string, regexp [, array])
Search string for the longest, leftmost substring matched by the regular expression regexp and return the character position (index) at which that substring begins (one, if it starts at the beginning of string). If no match is found, return zero..

edited Apr 25 '17 at 14:13

answered Apr 25 '17 at 13:33

RomanPerekhrest

88,541
4
65
105

Data that will be between "[" and "]" doesn't have to be at the 4th possition – Tadija Bagarić Apr 25 '17 at 13:35
1

$1 and $3 are mandatory? If data between `[` and `]` is at arbitrary position, could there be multiple entries like `aaa bbb ccc: ddd [fff] ggg hhh [jjj]` ? – RomanPerekhrest Apr 25 '17 at 13:36
Works great, tnx – Tadija Bagarić Apr 25 '17 at 13:50

dawg · Answer 2 · 2017-04-25T14:49:58.790

Given:

$ cat file
aaa bbb ccc: [DDD] efg oi    
aaa bbb [ccc:] lll DDD efg oo    
aaa [bbb] ccc: DDD

(note -- changed from the OP's example)

In POSIX awk:

awk 'BEGIN{fields[1]; fields[3]} 
                                {s=""
                                 for (i=1;i<=NF;i++) 
                                       if ($i~/^\[/ || i in fields) 
                                           s=i>1 ? s OFS $i : $i
                                 gsub(/\[|\]/,"",s)
                                 print s
                                }' file

Prints:

aaa ccc: DDD
aaa ccc:
aaa bbb ccc:

This does not print the field twice if it is both enclosed in [] and in the selected fields array. (i.e., [aaa] bbb ccc: does not print aaa twice) It will also print in correct field order if you have aaa [bbb] ccc ...

score -1 · Answer 3 · edited Apr 25 '17 at 20:23

-1

awk '$5=="[DDD]"{gsub("[\\[\\]]","");print $1,$3,$5}' file

or

awk '$5=="[DDD]"{print $1,$3, substr($5,2,3)}' file

aaa ccc: DDD

edited Apr 25 '17 at 20:23

Mad Physicist

107,652
25
181
264

answered Apr 25 '17 at 17:00

Claes Wikner

1,457
1
9
8

not sure, but code-only answers are generally frowned upon. – Jean-François Fabre Apr 25 '17 at 19:57

Bash include string between chars in awk print

3 Answers3