从命令输出的不同位置捕获子字符串

Question

I have a requirement, where I have to capture a string from a command's output and store it for further processing. Problem is that the command's output may change sometimes and hence this leads to erroneous results.

---

Requested dataset looks like

application_1532934978357_3376 app_name job_type user any_name_2 RUNNING 
UNDEFINED 10% hostname
application_1532934978357_3375 app_name job_type user any_name_2 RUNNING 
UNDEFINED 10% hostname
application_1532934978357_3374 app_name job_type user any_name_2 RUNNING 
UNDEFINED 10% hostname
application_1532934978357_249069 some_information_etc job_type any_name_2 
RUNNING UNDEFINED 95% hostname
application_1532934978357_239728 app_name job_type any_name_2 RUNNING 
UNDEFINED 10% hostname
application_1532934978357_89483 some_info job_type user any_name RUNNING 
UNDEFINED 10% hostname
application_1532934978357_248180 with prog_vrsn as
(se...select cast(Stage-27) job_type user any_name RUNNING UNDEFINED 36.1% 
hostname
application_15329349783879_657880 select cast
value ..(stage35) with table
where value=5; job_type user any_name RUNNING UNDEFINED 10% hostname

and I use:

cat in | grep "RUNNING" | grep "any_name" | awk '{print $1}'

which generates output as

application_1532934978357_89483 
(se...select cast(Stage-27)
where

While I want to produce output as :

application_1532934978357_89483 
application_1532934978357_248180 
application_15329349783879_657880 

Answer 1

Here is a GNU awk script that only captures the application_XXXX associated to the word any_name :

awk -v RS='[ \n]' '/application_[0-9_]+/{a=$0}/\<any_name\>/{print a}' file

It relies on the record separator RS that is set to capture each word. The application_XXXX string is stored in the variable a and printed when the word any_name is found.

Answer 2

You just need to add one more grep in your command:

command's output | grep "status_run" | grep -e "id_tag1" -e "id_tag2" | grep "app_id" | awk '{print $1}'

OR

awk '(/status_run/) && (/app_id*/) && (/id_tag[12]/) {print $1;}' filename

This will only print all the app_id with id_tag1 and id_tag2 and which has "status_run" in them.

---

Solution after updating your question:

cat filename | grep "RUNNING" | grep "any_name" | grep "application*" | awk '{print $1}'

If you want to print all the application Ids, then use the below command:

awk '/application*/{print $1}' filename