带有换行符的 sed 和 "答案

【问题标题】：sed with newlines and "带有换行符的 sed 和 "
【发布时间】：2012-04-25 14:51:42
【问题描述】：

我有一堆看起来像这样的文本文件：

His doctor attributed this to an allergy .

That hardly convinced him , as he had no history of allergies of any kind .
" Yet , that was to be the least of his problems .
I may have to take steroids for the rest of my life .
"
A topical steroid spray was later added to his repertoire of drugs and 
" he knew it was merely masking the underlying condition .
"

我想改变它，使. " 在一行中。所需的输出应如下所示：

His doctor attributed this to an allergy .

That hardly convinced him , as he had no history of allergies of any kind .
" Yet , that was to be the least of his problems .
I may have to take steroids for the rest of my life . "
A topical steroid spray was later added to his repertoire of drugs and 
" he knew it was merely masking the underlying condition . "

我试过了，但它不起作用：

sed -i 's/.\n"\n/. "\n/g'

有人可以帮助我使用正确的 sed 命令将 " 向上移动吗？

【问题讨论】：

它不起作用的原因是sed单独在行上工作。
那么有没有其他非sed的方式来解决呢？？
这将强制出现带有 " 的第二行。您不能明确查找换行符。现在，下一步是去掉引号并加入这两行。尝试替换时或搜索 \n，改用 $。这是连续处理两行的方法。我看到了 perl 解决方案，这可能是一个好方法。sed -e '/\.$/ { N /"/p }' test.txt N和/"/p}'之间有一个换行符。

标签： regex sed newline

【解决方案1】：

这是我想出来的：

sed -n '1{h;d};/^"$/{g;s/$/ "/p;n;h;d};x;p;${g;p}' input.txt

输出

His doctor attributed this to an allergy .

That hardly convinced him , as he had no history of allergies of any kind .
" Yet , that was to be the least of his problems .
I may have to take steroids for the rest of my life . "
A topical steroid spray was later added to his repertoire of drugs and
" he knew it was merely masking the underlying condition . "

【讨论】：

但我需要保留 \n 的其余部分，只有带有 "\n 的部分需要被截断。
感谢它创造了奇迹！！但为什么会这样呢？为什么需要这么复杂的正则表达式？
就像用汇编语言编程一样。您应该跟踪每一个微小的步骤。
h;d;g;x;p 是sed 命令。您可以输入man sed 以了解更多信息。
最后一个问题，我如何为一堆文件编写和编写它？以前我使用sed -i ... * 更改我的txtfiles 中的正则表达式。 -n 仅显示文本 -i，而您的正则表达式仅写入弄乱了 txtfiles。

【解决方案2】：

perl -00 -lpe 's/\n"$/"/mg'

产生所需的输出。

【讨论】：

【解决方案3】：

一个略有不同的sed 变体：

sed -n '1{h};1!{/"$/!H};/"$/{H;g;s/\.[ \n]*"$/\. "/;p;n;x}' input.txt

1 { h } — 将第一行放入保持缓冲区
1! { /"$/ !H } — 对于其余的行，如果没有孤独，则累积到保持缓冲区中 "
/"$/ { H; g; s/\.[ \n]*"$/\. "/; p; n; x } — 否则：
1. H — 添加到保持缓冲区
2. g — 将保持缓冲区移动到模式空间
3. s/\.[ \n]*"$/\. "/ — 替换
4. p — 打印出来
5. n — 阅读下一行
6. x——并将其保存在保留缓冲区中

【讨论】：

【解决方案4】：

这可能对你有用：

sed ':a;$!N;s/\.\n"/."/;P;D' /tmp/a
His doctor attributed this to an allergy .

That hardly convinced him , as he had no history of allergies of any kind ." Yet , that  was to be the least of his problems .
I may have to take steroids for the rest of my life ."
A topical steroid spray was later added to his repertoire of drugs and 
" he knew it was merely masking the underlying condition ."

【讨论】：