sed和regex打印仅匹配

Question

sed -er '.*(textsringhere.+?(?= ))(?:.*)((?>\d{4})-(?>\d{2})-(?>\d{2}) (?>\d{2}):(?>\d{2}):(?>\d{2})).* (ERROR.*)' errors.txt

Running the above command on my errors.txt file simply produces all the matching lines in that file. I had assumed that adding .* to the beginning of my regex would have forced sed to replace the entire line with the matches?

Sample

Input:

Aug  2 16:36:37 App.Dev.thing1 839854b7-749-4f12-89e0-3ad002ab5ffe[[APP/PROC/WEB/0]] 2017-08-02 16:36:37 [main] ERROR o.s.boot.SpringApplication - Application startup failed
Aug  2 18:04:46 App.Dev.thing2 eaedf253-df57-4c12-ade6-ea73274dbbc4[[APP/PROC/WEB/0]] 2017-08-02 18:04:46 [main] ERROR o.s.boot.SpringApplication - Application startup failed
Aug  3 01:45:55 App.Dev.thing2 eaedf253-df57-4c12-ade6-ea73274dbbc4[[APP/PROC/WEB/0]] 2017-08-03 01:45:55 [http-nio-8080-exec-1] ERROR c.c.v.b.m.c.i.thing2 - Error Processing Batch, this batch will be consumed individually`

Expected output:

App.Dev.thing1 ERROR o.s.boot.SpringApplication - Application startup failed
App.Dev.thing2 ERROR o.s.boot.SpringApplication - Application startup failed
App.Dev.thing2 ERROR c.c.v.b.m.c.i.thing2 - Error Processing Batch, this batch will be consumed individually`

Answer 1

Simple sed approach:

sed -E 's/.*(App\.\S+).+(ERROR .*)$/\1 \2/' errors.txt

The output:

App.Dev.thing1 ERROR o.s.boot.SpringApplication - Application startup failed
App.Dev.thing2 ERROR o.s.boot.SpringApplication - Application startup failed
App.Dev.thing2 ERROR c.c.v.b.m.c.i.thing2 - Error Processing Batch, this batch will be consumed individually