如何使用awk选择性处理文本行：保留指定#行并去#，移除无关#行

Let's get this sorted out for you! Your goal is to:

• Keep all lines that don't start with # exactly as they are
• For lines starting with #, only keep those containing Build or Type, and remove the leading # from those lines

The Right Awk Command #

Here's a concise, working command that does exactly what you need:

awk '/^#/ { if (/Build|Type/) { sub(/^#/, ""); print; next } } 1' input > output

Breakdown of How It Works #

Let's break down each part so you understand what's happening:

• /^#/: Targets lines that start with #
  • Inside this block: We check if the line contains either Build or Type with /Build|Type/
  • If it does, sub(/^#/, "") removes the leading #, we print the modified line, then next skips to the next line (so we don't process it again)
• 1: This is a shorthand in awk that evaluates to "true", which triggers the default action of printing the line. This handles all lines that don't start with #—they get printed as-is.

Why Your Previous Commands Didn't Work #

Let's quickly address the issues with your attempts:

• First command: awk '!/#/' input < out → This matches any line that doesn't contain # anywhere, so it deletes all lines with a # entirely—way too broad for your needs.
• Second command: awk '/#/ && !/Build|Length/' input < out → This matches lines that have a # and don't contain Build or Length, which is the opposite of what you want. That's why it left the unwanted #CN Filters: and # Flags = 1,2,3 lines.

Test with Your Example Input #

For your input:

#Build = NCBI Build 37
#CN Filters:
# Flags = 1,2,3
# Type = Lowess
Length
Event ID 1 Gain xxx
10 Loss yyy

Running the correct command will produce your desired output:

Build = NCBI Build 37
Type = Lowess
Length
Event ID 1 Gain xxx
10 Loss yyy

内容的提问来源于stack exchange，提问作者justaguy