sed例子

以care.log这个log文件为例，
care.log:

05:44:31,816 DEBUG RawAggregationWorker:70 - LTS is working on Raw Data Aggregation execute Start, Now:1535003071816, currentTimeMillis:1535003071816
05:44:31,816 DEBUG RawAggregationWorker:57 - LTS is working on Raw Data Aggregation to dataPoints:22, careConfig.getLtsRawaggregationDataPoints():10
05:46:53,264 DEBUG RawAggregationWorker:100 - LTS is working on Raw Data Aggregation execute End, Now:1535003071816, currentTimeMillis:1535003213264, now Done Took: 141448ms, currentTime Done Took: 141448ms
05:46:53,265 DEBUG MinAggregationWorker:59 - LTS is working on Min Data Aggregation execute Start, Now:1535002591816, currentTimeMillis:1535003213265
05:55:04,112 DEBUG MinAggregationWorker:108 - LTS is working on Min Data Aggregation execute End, Now:1535002591816, currentTimeMillis:1535003704112, now Done Took: 1112seconds, currentTime Done Took: 490seconds

05:55:04,112 DEBUG HourAgingWorker:53 - LTS is working on Hourly Aging, now: 1535003011816, currentTimeMillis:1535003704112
05:58:29,697 DEBUG HourAgingWorker:69 - LTS is working on Hourly Aging, now Done Took: 897881ms, currentTime Done Took: 205585ms

希望结果为：

Raw Data Aggregation Start-Now:1535003071816;currentTimeMillis:1535003071816;Raw Data Aggregation-dataPoints:10
Raw Data Aggregation End-Now:1535003071816;currentTimeMillis:1535003213264
Min Data Aggregation Start-Now:1535002591816;currentTimeMillis:1535003213265
Min Data Aggregation End-Now:1535002591816;currentTimeMillis:1535003704112
Hourly Aging-now: 1535003011816;currentTimeMillis:1535003704112
Hourly Aging-now Done Took: 897881ms;currentTime Done Took: 205585ms

匹配从行首到关键字并替换

sed 's#^.* on ##g'

匹配关键字并保留匹配到关键字其中一部分（用括号指代要保留的部分在#后用12.....表示）

sed 's#(^.*) execute (.*),.*(Now:[0-9]{13}), (currentTimeMillis:[0-9]{13}).*$#1 2-3;4#g'

合并匹配到的行与上一行

sed '{N;s/(.*)
(^.*dataPoints.*$)/1;2/;t merge;P;D};:merge;p;d'

删除空行

sed '/^$/d'

或

awk NF

最后，合并在一起：

 cat care.log | sed 's#^.* on ##g' | sed 's#(^.*) execute (.*),.*(Now:[0-9]{13}), (currentTimeMillis:[0-9]{13}).*$#1 2-3;4#g' | sed 's#^(.*) to (dataPoints).*DataPoints():([0-10]{1,})$#1-2:3#g' | sed 's#(^.*), (now.*), (current.*:.*)$#1-2;3#g' | sed '{N;s/(.*)
(^.*dataPoints.*$)/1;2/;t merge;P;D};:merge;p;d' |awk NF