Bash 如何使用sed在给定的行范围/模式内在匹配之前添加行？

我想做一个基本的

sed

-行插入，但要使用给定的行范围

使用给定的示例文件

sample.txt

：

foo
tomato
apple
bar
START_MARKER
tomato
apple
kiwifruit
pineapple
strawberry

END_MARKER

apple
qux

我想在包含“苹果”的“开始标记”和“结束标记”之间的任何一行之前添加一行（比如“\uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu。 我的样本将变成：

foo
tomato
apple
bar
START_MARKER
tomato
_____new_line_____
apple
kiwifruit
_____new_line_____
pineapple
strawberry

END_MARKER

apple
qux

我试过：

sed '/^START_MARKER$/,/^END_MARKER$/ /apple/i _____new_line_____' sample.txt

但我得到：

sed:-e表达式#1，字符33：未知命令：`/'

还尝试：

sed '/^START_MARKER$/,/^END_MARKER$/ {/apple/i _____new_line_____}' sample.txt

但我得到：

sed:-e表达式#1，字符0:不匹配的`{'

如果我想在包含“apple”的所有行（不使用开始标记和结束标记）之前添加一个“\uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu

sed '/apple/i _____new_line_____' sample.txt

---

我缺少将/start/、/end/模式选择与我的//I add line操作链接/组合起来的内容。

与

GNU sed

$ sed -e '/^START_MARKER$/,/^END_MARKER$/ { /apple/i _____new_line_____' -e '}' ip.txt
foo
tomato
apple
bar
START_MARKER
tomato
_____new_line_____
apple
kiwifruit
_____new_line_____
pineapple
strawberry

END_MARKER

apple
qux

a/c/i

命令之后的所有内容都被视为该命令的参数。因此，您需要使用换行符或如上所示的

-e

选项将它们分开。有关详细信息，请参阅

对于其他

sed

，这可能有效，但不确定，因为我无法测试它：

sed '/^START_MARKER$/,/^END_MARKER$/ {
/apple/i\
_____new_line_____
}' ip.txt

---

您可以使用以下选项：

sed '/START_MARKER/,/END_MARKER/!b;/apple/ i_____new_line_____' sample.txt > out

其中，

sed

遍历每一行，如果图案“apple”在开始标记和结束标记的范围内，则在“apple”之前插入“\uuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuuu

如果您不将输出定向到文件，则它将在终端屏幕上打印。

这可能适用于您（GNU-sed）：

这将在包含

apple

的

START

和

END

或文件结尾之间的任何行之前插入一行

要仅在包含

apple

的行之前的

START

和

END

之间插入一行，请使用：

sed -e ':a;/\n/{/^[^\n]*apple/i\__newline__' -e 'P;D};/START/{:b;N;/END/!bb;ba}' file

这将收集

START

和

END

之间的行，然后在包含

apple

的行之前插入一行

sed -e ':a;/\n/{/^[^\n]*apple/i\__newline__' -e 'P;D};/START/{:b;N;/END/!bb;ba}' file