很久没使用sed命令,今天有一个小需求需要对一批文件的内容进行修改,于是想起sed命令,顺便回顾一下。
文本的内容大致是
item [5]:
class = "IntervalTier"
name = "PhonAlign"
xmin = 0.0
xmax = 1.799688
intervals: size = 10
intervals [1]:
xmin = 0.0
xmax = 0.21
text = "#"
intervals [2]:
xmin = 0.21
xmax = 0.25
text = "ts"
intervals [3]:
xmin = 0.25
xmax = 0.28
text = "u"
intervals [4]:
xmin = 0.28
xmax = 0.51
text = "o"
intervals [5]:
xmin = 0.51
xmax = 0.71
text = "ts"
intervals [6]:
xmin = 0.71
xmax = 0.74
text = "u"
intervals [7]:
xmin = 0.74
xmax = 1.02
text = "o"
intervals [8]:
xmin = 1.02
xmax = 1.23
text = "ts"
intervals [9]:
xmin = 1.23
xmax = 1.76
text = "u"
intervals [10]:
xmin = 1.76
xmax = 1.799688
text = "o"
item [6]:
class = "IntervalTier"
name = "TokensAlign"
xmin = 0.0
xmax = 1.799688
intervals: size = 4
intervals [1]:
xmin = 0.0
xmax = 0.21
text = "#"
intervals [2]:
xmin = 0.21
xmax = 0.51
text = "做"
intervals [3]:
xmin = 0.51
xmax = 1.02
text = "做"
intervals [4]:
xmin = 1.02
xmax = 1.799688
text = "做"
item [7]:
class = "IntervalTier"
name = "Activity"
xmin = 0.0
xmax = 1.799688
intervals: size = 2
intervals [1]:
xmin = 0.0
xmax = 0.21
text = "silence"
intervals [2]:
xmin = 0.21
xmax = 1.799688
text = "speech"
复制代码
主要替换的位置是'做'这个次替换成对应的拼音
稍微观察一下文件的特点,由于替换的位置位于item[6]~item[7]这两个字符串之间。于是使用
echo "sed -i -E '/item \[6\]/,/item \[7\]/s/text = \"([^#\"]+)\"/text = \"${prex}\"/g' ${line}"|sh
复制代码
这里有一个小的知识点,使用sed如果直接用变量格式化,但没办法进行格式化。于是借助echo先将格式化好的字符串输出,然后调用sh命令执行。
另外还有一个用法,当需要在shell中判断某个字符串的包含关系,可以用如下的条件判断
if [[ $line =~ "_" ]]
复制代码