How can I use grep or any other utility to get a string from a structured data based on a condition?-CodePudding

Name                 State           Dns
DeltaService         running         DeltaService.test.qa.domain.com
DeltaService_1       stopped   
DeltaService_2       stopped         DeltaService_2.test.qa.domain.com
UnitedService        running         UnitedService.test.qa.domain.com
UnitedService_1      stopped         
UnitedService_2      running         UnitedService_2.test.qa.domain.com

The above data shows as part of my logs during execution. What I want to do is to search the logs for services with a similar name that have a DNS and store them in an array using bash script. For example if I want DeltaService, I want to be able to retrieve "DeltaService" and "DeltaService_2" and store in an array where I can use those service names for further processing. How can I achieve this? What I have tried:

#"${serviceLogs[@]}" = contains the structured data.

print -- '%s\n' "${serviceLogs[@]}" | grep -oh "\w*DeltaService\w*"

output:
DeltaService
DeltaService_1
DeltaService_2

#I don't need DeltaService_1 because it has no DNS.

CodePudding user response：

awk seems to be the right tool for that:

printf -- '%s\n' "${serviceLogs[@]}" |
awk '$1 ~ /^DeltaService/ && $3 != "" {print $1}'

DeltaService
DeltaService_2

CodePudding user response：

Per OP's comments:

serviceLogs='Name                 State           Dns
DeltaService         running         DeltaService.test.qa.domain.com
DeltaService_1       stopped   
DeltaService_2       stopped         DeltaService_2.test.qa.domain.com
UnitedService        running         UnitedService.test.qa.domain.com
UnitedService_1      stopped         
UnitedService_2      running         UnitedService_2.test.qa.domain.com'

One idea using awk to parse the data:

$ awk -v ptn="DeltaService" '$1~ptn && $3 {print $1}' <<< "${serviceLogs}"
DeltaService
DeltaService_2

Storing the results in a bash array named services[]:

$ services=( $(awk -v ptn="DeltaService" '$1~ptn && $3 {print $1}' <<< "${serviceLogs}") )
$ typeset -p services
declare -a services=([0]="DeltaService" [1]="DeltaService_2")