CODEWITHSUNDEEP
bashunixsedawk
C Dubya
C Dubya

Reputation: 135

Remove text from line

I have a log file that I essentially cat the entire log out, cut it up strings until there are only two fields remaining, sort uniquely on field one and sum up field to the right when field numbers and field1 are the same. Example

80 128 
443 40
80 100
25  20 
443 44
80 128

The results would be 

80 356
443 84
25  20

The issue I am having is there is an inconsistency in the first field that I cut out where sometimes output will look like:

80 128

and sometimes it is 

80(LOCAL\randomuser) 128

So my output ends up looking like

80 356
80(LOCAL\randomuser) 128
443 84
25  20

this is because 80(LOCAL\randomuser) is a unique line.

How can I first normalize the first field so that (LOCAL\randomuser) is removed from lines where exist while lines that do not have (LOCAL\randomuser) remain the same.

Upvotes: 0

Views: 151

Answers (3)

Scrutinizer
Scrutinizer

Reputation: 9936

Or force $1 into numerical context:

awk '{A[$1+0]+=$NF} END{for (i in A) print i, A[i]}' file

Upvotes: 1

Fredrik Pihl
Fredrik Pihl

Reputation: 45670

Use

awk '{a[$1]+=$NF} END{for (i in a) print i, a[i]}' input

I.e use first field as key and add the last field.

If there isn't a space between the first number and the ( as it looks like in your example, tell awk to split on ( as well:

awk -F"[ (]+" '{a[$1]+=$NF} END{for (i in a) print i, a[i]}' input

output:

$ awk '{a[$1]+=$NF} END{for (i in a) print i, a[i]}' input
25 20
80 356
443 84

Another approach to remove the just the (LOCAL\randomuser) if present, you can use sed:

sed 's/(.*)//' input

Upvotes: 1

salezica
salezica

Reputation: 77089

grep -v will match lines that do not contain a pattern. Pipe the output through grep before it reaches cut.

Upvotes: 0

Related Questions