如何结合文本限定符使用cut命令提取CSV指定列

Ah, I feel your pain—cut is super handy for simple delimited data, but it completely ignores CSV's quoted field rules. That's why it's chopping your "1,2,3" into just "1" when you use -d',' -f2.

The good news is there are several reliable ways to handle this properly. Here are the most practical methods:

1. Use a CSV-aware awk script #

awk can be modified to track whether we're inside a quoted field, so it skips commas within quotes. Here's a one-liner that does exactly what you need:

awk -F',' '{
    in_quote = 0
    current_col = ""
    for (i=1; i<=NF; i++) {
        current_col = current_col $i
        # Toggle quote state if this field has an odd number of quotes
        if (gsub(/"/, "", $i) % 2 != 0) {
            in_quote = !in_quote
        }
        # If we're not in quotes and we've reached column 2, print and exit loop
        if (!in_quote && i == 2) {
            print current_col
            break
        }
        # If we're still in quotes, add back the comma we split on
        else if (in_quote) {
            current_col = current_col ","
        }
    }
}' xyz.txt

This will output "1,2,3" (keeping the quotes). If you want to strip the quotes, change the print line to:

print gensub(/^"|"$/, "", "g", current_col)

2. Use csvkit (the easiest dedicated CSV tool) #

If you can install external tools, csvkit is built for exactly this scenario. Install it via your package manager:

• Debian/Ubuntu: sudo apt install csvkit
• macOS: brew install csvkit
• Windows: pip install csvkit

Then extract the second column with:

csvcut -c 2 xyz.txt

By default, this outputs 1,2,3 (without quotes). To keep the quotes, add the --keep-quotes flag:

csvcut -c 2 --keep-quotes xyz.txt

3. Quick sed hack (for simple cases) #

If your CSV is as straightforward as your example, a sed command can work—though it's less robust for edge cases (like escaped quotes inside fields):

# Keep quotes
sed -E 's/^[^,]*,("[^"]*").*/\1/' xyz.txt

# Remove quotes
sed -E 's/^[^,]*,\"([^\"]*)\"(.*)/\1/' xyz.txt

Why cut fails here #

cut treats every comma as a separator, no exceptions. So when it processes A,"1,2,3",C,D, it splits the line into these fields:

1. A
2. "1
3. 2
4. 3"
5. C
6. D
   That's why -f2 gives you "1" instead of the full quoted field.

内容的提问来源于stack exchange，提问作者P S