CODEWITHSUNDEEP
bashgrep
Sasank Sai Sujan Adapa
Sasank Sai Sujan Adapa

Reputation: 43

how to collect maximum and minimum values from a grep search pattern using bash?

I have a huge log.txt file from which I need to calculate the maximum and minimum values.

 0-00:42:35.598 <tc_testcase>:[DEFAULT]:[PRINT]: tc_testcase.c:8963: VERIFY_CASE: range: X_PER_Y_USER_RANGE_0[1](a, 1)=121, (0)=0, (25000)=25000, (ok)
 0-00:42:35.598 <tc_testcase>:[DEFAULT]:[PRINT]: tc_testcase.c:8963: VERIFY_CASE: range: X_PER_Y_USER_RANGE_0[1](a, 1)=256879, (0)=0, (25000)=25000, (ok)
 0-00:42:35.598 <tc_testcase>:[DEFAULT]:[PRINT]: tc_testcase.c:8963: VERIFY_CASE: range: X_PER_Y_USER_RANGE_0[1](a, 1)=2300, (0)=0, (25000)=25000, (ok)
 0-00:42:35.598 <tc_testcase>:[DEFAULT]:[PRINT]: tc_testcase.c:8963: VERIFY_CASE: range: X_PER_Y_USER_RANGE_0[1](a, 1)=56897132, (0)=0, (25000)=25000, (ok)
 0-00:42:35.598 <tc_testcase>:[DEFAULT]:[PRINT]: tc_testcase.c:8963: VERIFY_CASE: range: X_PER_Y_USER_RANGE_0[1](a, 1)=12579, (0)=0, (25000)=25000, (ok)
 0-00:42:35.598 <tc_testcase>:[DEFAULT]:[PRINT]: tc_testcase.c:8963: VERIFY_CASE: range: X_PER_Y_USER_RANGE_0[1](a, 1)=968746, (0)=0, (25000)=25000, (ok)

so, as the first step I started to collect lines as below into a separate file with the below grep command and later to sort it:

 grep -Po '.X_PER_Y_USER_RANGE_0.*[1].*(a, 1).\K.*' file.txt > collect.txt

But the output I receive is quite different. It looks something like this:-

=121, (25000)=25000, (ok)
=256879, (25000)=25000, (ok)
=2300, (25000)=25000, (ok)
=56897132, (25000)=25000, (ok)
=12579, (25000)=25000, (ok)
=968746, (25000)=25000, (ok)

Expected should be:-

121
256879
2300
56897132
12579
968746

Can anyone help to modify the current grep command i'm using to collect the values as the expected output

Upvotes: 3

Views: 1010

Answers (2)

Shakiba Moshiri
Shakiba Moshiri

Reputation: 23864

You just needed to escape the special regex characters :

grep -Po 'range: X_PER_Y_USER_RANGE_0\[1\]\(a, \d\)=\K\d+' file.txt
121
256879
2300
56897132
12579
968746

For such a case (to me) Perl itself is preferable :

perl -lne '/ \d\)=\K\d+/ && print $&' file.txt
121
256879
2300
56897132
12579
968746

and for min and max something like this:

perl -lne '/ \d\)=\K\d+/g && push(@number,$&); END{ print "@{[sort {$a <=> $b} @number]}[0,-1]"}' file.txt
121 56897132

and END { ... } means sort them and take out 0 and last index of the array of @number

For sort using and not Perl to have min and max you can pipe your output to:

... | { mapfile -t arr; paste <(sort -n <(tr ' ' '\n' <<<  ${arr[@]})) <(sort -rn <(tr ' ' '\n' <<<  ${arr[@]})); }
121 56897132
2300    968746
12579   256879
256879  12579
968746  2300
56897132    121

which creates an array and then sort it twice. Then just pipe it to head -n 1 e.g.

... | head -n 1
121 56897132

Upvotes: 2

RavinderSingh13
RavinderSingh13

Reputation: 133620

You were close, trying to fix OP's attempt here. This could be done in GNU grep like following. We could use GNU grep's -P option to enable PCRE regex there.

grep -oP '.*range: X_PER_Y_USER_RANGE_0\[1\]\([a-zA-Z]+, \d+\)=\K\d+'  Input_file

Explanation: Simple explanation would be, firstly enabling -oP options for GNU grep to enable PCRE regex capabilities and to get only matched values. Then in main program of grep, using regex .*range: X_PER_Y_USER_RANGE_0\[1\]\([a-zA-Z]+, \d+\)= to match everything from starting to till (), note here escaping part of [, ],(,) to make those characters as literal ones. Then using \K option to forget all the matched values till now. Mentioning \d+ will match digits after it which is OP's actual requirement to get the digits.

Upvotes: 6

Related Questions