通过shell处理多行数据

### 源文件
cat > tmpb <<'EOF'
dbname:db_a,start_time:::
query_end_time:::,query_total_time:,data_record_num is [],data_size is 337K
dbname:db_a,end_time:::
total time:
==============================
dbname:db_b,start_time:::
query_end_time:::,query_total_time:,data_record_num is [],data_size is 20M
dbname:db_b,end_time:::
total time:
==============================
dbname:db_c,start_time:::
query_end_time:::,query_total_time:,data_record_num is [],data_size is 497K
dbname:db_c,end_time:::
total time:
==============================
dbname:db_d,start_time:::
query_end_time:::,query_total_time:,data_record_num is [],data_size is .4K
dbname:db_d,end_time:::
total time:
==============================
EOF

### 清洗数据
cat tmpb|while read line
do
    if [ "${line}" == '==============================' ]
    then
        echo ""
    else
        echo -n "${line}|"
    fi
done \
|awk -F'[|:, ]' '{print $2"|"$4":"$5":"$6"|"$8":"$9":"$10"|"$12"|"$15"|"$18"|"$27}'|sed 's#\[##g'|sed 's#]##g'|sort -t '|' -nk5