使用awk的gsub多模式匹配替换仅生效一个模式的问题

问题场景 #

执行awk脚本进行多模式搜索替换时，只有第一个替换规则生效，后续规则被忽略：
原执行命令：

output=$(echo "$output4" | awk 'NR>1 {gsub(/\[|]/,""); gsub("+r","READ"); gsub("+r+w","READ and WRITE"); gsub("+r+w+g","READ, WRITE and GRANT"); print}')

实际输出：

Group Analyst_Corp in security domain Native: READ. Group Finance_Dev in security domain Native: READ+w. Manufacturing_Corp in security domain Native: READ+w
Information_Tech in security domain Native: READ+w+g.

期望输出：

Group Analyst_Corp in security domain Native: READ. Group Finance_Dev in security domain Native: READ and WRITE. Manufacturing_Corp in security domain Native: READ and WRITE
    Information_Tech in security domain Native: READ, WRITE and GRANT.

问题原因 #

替换顺序搞反了：先执行gsub("+r","READ")会把所有+r替换成READ，导致原本的+r+w变成READ+w，+r+w+g变成READ+w+g，后面的规则根本匹配不到目标字符串，自然不会生效。

解决办法 #

把替换顺序改成从最长的模式到最短的模式，先处理最长的+r+w+g，再处理+r+w，最后处理+r，这样长模式不会被提前替换破坏：

修正后的命令：

output=$(echo "$output4" | awk 'NR>1 {gsub(/\[|]/,""); gsub("+r+w+g","READ, WRITE and GRANT"); gsub("+r+w","READ and WRITE"); gsub("+r","READ"); print}')

另外注意：+在正则里是特殊字符，最好用反斜杠转义，写成\+避免意外匹配，更严谨的写法：

output=$(echo "$output4" | awk 'NR>1 {gsub(/\[|]/,""); gsub(/\+r\+w\+g/,"READ, WRITE and GRANT"); gsub(/\+r\+w/,"READ and WRITE"); gsub(/\+r/,"READ"); print}')

这样执行后就能得到期望的输出结果。

内容的提问来源于stack exchange，提问作者bob freeman