在Perl中，您不能捕获可选组。但是您可以使用包含捕获组的非捕获可选组(?:...)?。

#!/bin/bash
strings=(
    'Execution error for request 78471. Reason: ESS-07033 Job logic indicated a system error occurred while executing an asynchronous java job for request 78471. Job error is: JBO-26067:'
    'Execution error for request 78471. Reason: ESS-07033 blah blah'
    'Execution error for request 78471. Reason: null'
)

for s in "${strings[@]}" ; do
    perl -lne '($req, $rea, $job) = /request\ ([0-9]+)
                                    (?:.*
                                        (ESS-[0-9]+)
                                    )?
                                    (?:.*
                                        (JBO-[0-9]+)
                                    )?
                                    /x;
               print "$req|$rea|$job"' <<< "$s"
done

字符串
我使用了/x修饰符，这使得正则表达式的格式更易读（它忽略了非引号的空格）。

$ sed -E 's/.* request ([0-9]+).*Reason: *(ESS-[0-9]+)?.* (JBO-[0-9]+)?.*/\1|\2|\3/g' file
78471|ESS-07033|JBO-26067
78471|ESS-07033|
78471||

$ cat file
Execution error for request 78471. Reason: ESS-07033 Job logic indicated a system error occurred while executing an asynchronous java job for request 78471. Job error is: JBO-26067:
Execution error for request 78471. Reason: ESS-07033 blah blah
Execution error for request 78471. Reason: null

字符串

这可能是你想要的，使用任何awk：

$ cat tst.awk
BEGIN { OFS="|" }
{
    req = ( match($0,/request /) ? substr($0,RSTART+RLENGTH) + 0 : "" )
    ess = ( match($0,/ESS-[0-9]+/) ? substr($0,RSTART,RLENGTH) : "" )
    jbo = ( match($0,/JBO-[0-9]+/) ? substr($0,RSTART,RLENGTH) : "" )
    print req, ess, jbo
}

字符串

$ awk -f tst.awk file
78471|ESS-07033|JBO-26067
78471|ESS-07033|
78471||

型
并借用创建的示例输入@Choroba：

$ cat file
Execution error for request 78471. Reason: ESS-07033 Job logic indicated a system error occurred while executing an asynchronous java job for request 78471. Job error is: JBO-26067:
Execution error for request 78471. Reason: ESS-07033 blah blah
Execution error for request 78471. Reason: null

型