在Linux中删除包含字符串的多个文件中的一行

Question

I am moving some of my pages from 8859 to UTF so I just included the UTF in my file that is included on most pages. I am trying to erase the line with the charset definition from pages with that include.

find . -type f -exec grep -lr 'headIncluded' {} + -exec sed -i '/meta http-equiv=\\"content-type\\" content=\\"text\\/html; charset/d' {} \\;

I thought this one should work, but I noticed that it erased the line in a few pages where headIncluded was not present. Any suggestions to what is wrong with this command?

Answer 1

为什么不尝试：

grep -lr "headIncluded" /pathtodirectory/* | xargs sed -i '/meta http-equiv=\"content-type\" content=\"text\/html; charset/d'

Answer 2

When you use + to terminate the -exec , the command is run with multiple files. If any of those files contain the string, the grep will succeed. Also, from the documentation "this variant of -exec always returns true", so the return value of grep is irrelevant.