CODEWITHSUNDEEP
linuxbashshellunixawk
Moshe Shmukler
Moshe Shmukler

Reputation: 1300

awk parsing output and extracting values

I need to process output that looks as below:

my_first_key: {i}text
my_first_key: {j}different_text
my_first_key: {k}some_text
my_second_key: value1
my_first_key: {l}some_text
my_second_key: value2

Lines always begin with one of two keys followed by : then a space then either an index in curly braces for one key, or immediately a value for the second key. A line with the second-type key and the value is always preceded by a line with first key and the index.

I need two functions:

  1. array(): to return the array as {k:value1, l:value2}.
  2. index(value) to return the index value from the preceding line so index(value2) returns l.

I see examples to return line numbers etc., but my awk level is 0.

[From Moshe comment] Algorithm is like this:

a) awk '/$value/{ print NR; exit }'
b) awk 'NR==$(previous_return - 1)
c) awk '{split($0,a,"{}") | print a[2]

Upvotes: 0

Views: 848

Answers (1)

r.g
r.g

Reputation: 56

I don't have too much time for now, but the following program should help you solve the first point. It is strictly based on your example. From your example input, I replaced alphabetic indexes i,j,k and l with 1,2,3,4, so the test file contains :

*my_first_key: {1}text
my_first_key: {2}different_text
my_first_key: {3}some_text
my_second_key: value1
my_first_key: {4}some_text
my_second_key: value2*

and the program is :

BEGIN {
  # split on braces to have the index from my_first_key: in field #2 without effort
  FS="[{}]"
}
/^my_first_key:/ { ix=$2}
/^my_second_key:/ {
  sub(/^my_second_key: */, "") # leaves only value2 in $0
  if (first_done==0) {
    buffer["index1"]=ix
    buffer["value1"]= $0
    first_done=1
  }
  else { # first_done == 1
    printf ("{%s:%s, %s:%s}\n", buffer["index1"], buffer["value1"], ix, $0)
    first_done=0
  }
}

and the output is:

awk -f example.awk test.data
{3:value1, 4:value2}

Upvotes: 3

Related Questions