如何替换出现的管道 (|)，而不替换用作分隔符的管道

Question

在 Linux 服务器上，我有一些以竖线分隔的文件，这些文件在某些字符串列中具有竖线 (|) 字符。这些文件是用管道分隔的，但分隔符管道始终用双引号引起来，而文本/字符串中的管道可以在其周围包含任何字符、空格等。我想用破折号 (-) 替换字符串中的非分隔符管道

这是一个例子：

col1|col2|col3
"1"|"This is my column 2 |Although there is pipe here, it is not a delimiter pipe."|"And this is my 2nd column '|" with a pipe followed by a double quote"|"|and finally this 3rd column starts with a | that is not'|a delimiter

我尝试了一些 SED 和 AWK 命令，但无法从字符串的替换中正确排除分隔符（“|”）。

我正在尝试获得这样的输出：

col1|col2|col3
"1"|"This is my column 2 -Although there is pipe here, it is not a delimiter pipe."|"And this is my 2nd column '-" with a pipe followed by a double quote"|"-and finally this 3rd column starts with a - that is not'-a delimiter

Answer 1

我能想到的方法之一就是替换“|”用一些特殊字符（在本例中为##）并替换所有出现的 |用 - 然后将 ## 替换为“|”

cat file.txt | sed -e "s/\"|\"/##/g" | tr "|" - | sed -e "s/##/\"|\"/g"

输出-

col1-col2-col3
"1"|"This is my column 2 -Although there is pipe here, it is not a delimiter pipe."|"And this is my 2nd column '-" with a pipe followed by a double quote"|"-and finally this 3rd column starts with a - that is not'-a delimiter