在匹配的块之前插入代码块。

huangapple
go评论86阅读模式
英文:

SED INSERT block of code BEFORE matched block

问题

INSERT BEFORE:

  1. sed '/Categ         #n/{N;N;s/Categ         #n\nItemName      1/&\n**new_line1**\n**new_line2**\n**new_line3**/}' input.txt

APPEND AFTER:

  1. sed '/ItemName      2/{N;s/\(ItemName      2\)\(.*\)/\n**new_line1**\n**new_line2**\n**new_line3**/}' input.txt
英文:

I have a text file I want to edit during a script excecution.
I want to pass from the text on left to the text on right using sed insert:

INSERT BEFORE:
from:

  1. [...]
  2. [ placeholder ]
  3. Categ         #n
  4. ItemName      1
  5. [...]
  6. [ placeholder ]
  7. Categ         #n
  8. ItemName      2
  9. [...]

to:

  1. [...]
  2. [ placeholder ]
  3. Categ         #n
  4. ItemName      1
  5. [...]
  6. **new_line1**        <<<--- INSERT BEFORE
  7. **new_line2**        |||
  8. **new_line3**        |||
  9. [ placeholder ]      <<<--- 3-lines match
  10. Categ         #n     |||
  11. ItemName      2      |||
  12. [...]

I have tried to adapt sed append lines to sed insert.
The appending works, the insert does not.
The match to append was 2 lines one after the other.
In the insert, the match it is 3 lines and I tried to match the first and the last, as I don't know how to match the entire 3-lines block.

APPEND AFTER:

  1. sed -e '/^\[ placeholder \]\/ItemName      2/a \\nnew_line1\nnew_line2\nnew_line3' input.txt

and it works from:

  1. [...]
  2. [ placeholder ]
  3. ItemName      1
  4. [...]
  5. [ placeholder ]
  6. ItemName      2
  7. [...]

to:

  1. [...]
  2. [ placeholder ]
  3. ItemName      1
  4. [...]
  5. [ placeholder ]     <<<--- 2-lines match
  6. ItemName      2     |||
  7. **new_line1**       <<<--- APPENDED AFTER
  8. **new_line2**       |||
  9. **new_line3**       |||
  10. [...]

I have tried to use the syntax examples in this post, but with no success.
The sed - Stream Editor

答案1

得分: 0

问题会变得简单得多，如果文件是倒着读的。

可以使用sed解决方案，但awk非常直观：

  1. tac input.txt |
  2. awk '
  3.     # 将每一行输入复制到输出
  4.     { print }
  6.     # 查找所需的项目名称行
  7.     #   如果找到，改变状态并跳过剩余的测试（无法匹配）
  8.     $1=="ItemName" && $2==2 { s = 1; next }
  10.     # 如果我们到达这里，查找占位符行
  11.     #   如果找到，追加新行并重置状态
  12.     s && /\[ placeholder \]/ {
  13.         print "new3"
  14.         print "new2"
  15.         print "new1"
  16.         s = 0
  17.     }
  18. ' | tac > output.txt
英文:

The problem becomes much simpler if the file is read backwards.

A sed solution is possible but awk is very straightforward:

  1. tac input.txt |
  2. awk '
  3.     # copy every input line to output
  4.     { print }
  6.     # look for desired itemname line
  7.     #   if found, change state and
  8.     #   skip remaining tests (which can't match)
  9.     $1=="ItemName" && $2==2 { s = 1; next }
  11.     # if we get here, look for placeholder line
  12.     #   if found, append new lines and reset state
  13.     s && /\[ placeholder \]/ {
  14.         print "new3"
  15.         print "new2"
  16.         print "new1"
  17.         s = 0
  18.     }
  19. ' | tac > output.txt

答案2

得分: 0

如果您的输入中不包含NUL字节，您可以尝试使用GNU sed执行以下操作：

  1. $ from='[ 占位符 ]\nCateg         #n\nItemName      2\n'
  2. $ to='**新行1**\n**新行2**\n**新行3**\n'
  3. $ sed -Ez "s/(^|\n)($from)/$to/" input.txt
  4. [...]
  5. [ 占位符 ]
  6. Categ         #n
  7. ItemName      1
  8. [...]
  9. **新行1**
  10. **新行2**
  11. **新行3**
  12. [ 占位符 ]
  13. Categ         #n
  14. ItemName      2
  15. [...]

使用-z选项将输入视为一组以零字节（ASCII NUL字符）终止的行，而不是以换行符终止。如果输入不包含NUL字节，这将允许一次性处理整个文件。当然，您必须正确转义from和to中的特殊字符。

英文:

If your input does not contain NUL bytes you could try the following with GNU sed:

  1. $ from='\[ placeholder \]\nCateg         #n\nItemName      2\n'
  2. $ to='**new_line1**\n**new_line2**\n**new_line3**\n'
  3. $ sed -Ez "s/(^|\n)($from)/$to/" input.txt
  4. [...]
  5. [ placeholder ]
  6. Categ         #n
  7. ItemName      1
  8. [...]
  9. **new_line1**
  10. **new_line2**
  11. **new_line3**
  12. [ placeholder ]
  13. Categ         #n
  14. ItemName      2
  15. [...]

With the -z option the input is treated as a set of lines, each terminated by a zero byte (the ASCII NUL character) instead of a newline. If the input does not contain NUL bytes this allows to process the entire file at once. Of course you must properly escape special characters in from and to.

答案3

得分: 0

以下是已翻译的代码部分：

  1. A loop probably makes sense to accumulate the block into the pattern space:
  3.     sed '/placeholder/{:l /ItemName/!{N;bl};/ItemName.*2/s/^/new1\nnew2\nnew3\n/}' file 
  5. Note that if `ItemName` is always two lines after `placeholder` then this simpler method also works:
  7.     sed '/placeholder/{N;N;/ItemName.*2/s/^/new1\nnew2\nnew3\n/}' file
  9. There is also this complicated method of accumulating in the hold space rather than the pattern space:
  11.     sed '/placeholder/,/ItemName/{/placeholder/h;//!H;/ItemName/!d;//g;/ItemName.*2/s/^/new1\nnew2\nnew3\n/}' file
  13. All methods produce this output:
  15.     [...]
  16.     [ placeholder ]
  17.     Categ         #n
  18.     ItemName      1
  19.     [...]
  20.     new1
  21.     new2
  22.     new3
  23.     [ placeholder ]
  24.     Categ         #n
  25.     ItemName      2
  26.     [...]

这是你提供的代码的翻译部分。

英文:

A loop probably makes sense to accumulate the block into the pattern space:

  1. sed '/placeholder/{:l /ItemName/!{N;bl};/ItemName.*2/s/^/new1\nnew2\nnew3\n/}' file

Note that if ItemName is always two lines after placeholder then this simpler method also works:

  1. sed '/placeholder/{N;N;/ItemName.*2/s/^/new1\nnew2\nnew3\n/}' file

There is also this complicated method of accumulating in the hold space rather than the pattern space:

  1. sed '/placeholder/,/ItemName/{/placeholder/h;//!H;/ItemName/!d;//g;/ItemName.*2/s/^/new1\nnew2\nnew3\n/}' file

All methods produce this output:

  1. [...]
  2. [ placeholder ]
  3. Categ         #n
  4. ItemName      1
  5. [...]
  6. new1
  7. new2
  8. new3
  9. [ placeholder ]
  10. Categ         #n
  11. ItemName      2
  12. [...]

The simpler method is simply using N to append more lines into the pattern space before it's printed.

The loop adds a label :l and b branches to that label until ItemName.

The hold space method uses h and H to hold, d to delete and not print, g to get. As well as // as shorthand to repeat same match.

I have been lazy with the matches and the lines inserted to make this all more readable. ItemName.*2 is especially lazy. Probably should be something like ItemName\s\+2\b

答案4

得分: 0

这可能适用于您（使用GNU sed和bash）：

  1. match='[ placeholder ]
  2. Categ         #n
  3. ItemName      2'
  4. insert='**new_line1**\n**new_line2**\n**new_line3**'
  5. sed '1N;N;/^'"${match}"'$/!{P;D};i\'"${insert}"' file
  6. 打开一个三行窗口，直到找到匹配项，然后插入所需的字符串，之后两行窗口将继续，直到文件末尾。
英文:

This might work for you (GNU sed and bash):

  1. match=''\[ placeholder \]\nCateg         #n\nItemName      2'
  2. insert='**new_line1**\n**new_line2**\n**new_line3**'
  3. sed '1N;N;/^'"${match}"'$/!{P;D};i\'"${insert}" file

Open a three line window until a match occurs, then insert the required string and thereafter the two line window will continue until the end of file.

huangapple
  • 本文由 发表于 2023年7月13日 18:10:56
  • 转载请务必保留本文链接：https://go.coder-hub.com/76678234.html
  • append
  • codeblocks
  • insert
  • match
  • sed
匿名

发表评论

匿名网友

:?: :razz: :sad: :evil: :!: :smile: :oops: :grin: :eek: :shock: :???: :cool: :lol: :mad: :twisted: :roll: :wink: :idea: :arrow: :neutral: :cry: :mrgreen:

确定