Shell gets two columns of values from a single line of files

The contents of the log file are as follows:
logid=2210608107 worker_id=140610281404160 optime=1452598610.631276 msg=service+not+found+in+share+memory service=Redis_doc_push

The post-column output is obtained through awk, and the disadvantage is that if the log is printed in a format that does not have spaces, you cannot use the
cat xxx.log | grep -oP "logid=\d+.*service=\w+" | awk -F‘ ‘ ‘{print $1,$NF}‘

Capturing a group with a regular expression, replacing the entire row with a value of two groups with SED
cat xxx.log | sed -r "s/.*?logid=([0-9]+).*?service=(\w+).*?/\1,\2/"

Shell gets two columns of values from a single line of files