尝试在bash / linux中删除文件符合目标条件的文件中以特定字符串开头的行

I am writing a bash script that will run a couple of times a minute. What I would like it to do is find all files in a specified directory that contain a specified string, and search that list of files and delete any line beginning with a different specific string (in this case it's

我正在编写一个每分钟运行几次的bash脚本。我想要它做的是找到指定目录中包含指定字符串的所有文件，并搜索该文件列表并删除以不同特定字符串开头的任何行（在这种情况下，它是

Here's what I've tried s far, but they aren't working:

这是我尝试过的，但它们不起作用：

ls -1t /the/directory | head -10 | grep -l "qualifying string" * | sed -i '/^<meta/d' *'
ls -1t /the/directory | head -10 | grep -l "qualifying string" * | sed -i '/^<meta/d' /the/directory'

The only reason I added in the head -10 is so that every time the script runs, it will start by only looking at the 10 most recent files. I don't want it to spend a lot of time searching needlessly through the entire directory since it will be going through and removing the line many times a minute.

我在头部-10中添加的唯一原因是，每次脚本运行时，它将仅从查看最近的10个文件开始。我不希望它花费大量时间在整个目录中进行不必要的搜索，因为它将通过并每分钟删除一次。

The script has to be run out of a different directory than the files are in. Also, would the modified date on the files change if the "<meta" string doesn't exist in the file?

该脚本必须运行在与文件不同的目录中。此外，如果文件中不存在“ ”字符串，文件上的修改日期是否会更改？

2 个解决方案

#1

There are a variety of problem with this part of the command...

这部分命令存在各种问题......

ls -1t /the/directory | head -10 | grep -l "qualifying string" * ...

First, you appear to be trying to pipe the output of ls ... | head -10 into grep, which would cause grep to search for "qualifying string" in the output of ls. Except then you turn around and provide * as a command line argument to grep, causing it to search in all the files, and completely ignoring the ls and head commands.

首先，您似乎试图管道ls的输出... |将-10转入grep，这会导致grep在ls的输出中搜索“qualifying string”。除此之外，您转过来并将*作为命令行参数提供给grep，使其搜索所有文件，并完全忽略ls和head命令。

You probably want to read about the xargs commands, which reads a list of files on stdin and then runs a given command against that list. For example, you ought to be able to generate your file list like this:

您可能想要阅读有关xargs命令的信息，该命令读取stdin上的文件列表，然后针对该列表运行给定命令。例如，您应该能够像这样生成文件列表：

ls -1t /the/directory | head -10 | 
  xargs grep -l "qualifying string"

And to apply sed to those files:

并将sed应用于这些文件：

ls -1t /the/directory | head -10 | 
  xargs grep -l "qualifying string" |
  sed -i 's/something/else/g'

Modifying the files with sed will update the modification time on the files.

使用sed修改文件将更新文件的修改时间。

#2

You can use globbing with the * character to expand file names and loop through the directory.

您可以使用带有*字符的globbing来扩展文件名并循环遍历目录。

n=0
for file in /the/directory/*; do
    if [ -f "$file" ]; then
        grep "qualifying string" "$file" && sed -i '/^<meta/d' "$file"
        n=$((n+1))
    fi
    [ $n -eq 10 ] && break
done

#1