关于sort工具-k选项中字母排序参数的详细解释请求

Let's walk through each single-letter flag you can pair with the -k (key) option in sort. These flags let you override global sort settings only for the specific key you're defining—super handy when you need different sorting rules for different columns or fields in your data.

• b: Ignore leading blanks (whitespace) when comparing the key. If your data has fields with leading spaces, this ensures sort doesn’t count those spaces as part of the value for that key.
  Example: Using -k1,1b will sort 123 and 45 as 45 before 123 (since it skips the leading spaces).

• d: Use dictionary order—only consider letters, digits, and whitespace. Special characters (like !@#$) get ignored during the key comparison. Ideal for sorting text where punctuation doesn’t matter.

• f: Fold lowercase to uppercase (case-insensitive sorting) for the key. So apple and Apple will be treated as identical for this key, even if your global sort settings are case-sensitive.

• g: Use general numeric value sorting. This handles floating-point numbers and scientific notation (like 1.2e3) correctly—more flexible than the integer-focused -n flag.

• i: Ignore non-printable characters when comparing the key. If your data has hidden control characters, this flag makes sort skip them for the specific key you’re configuring.

• M: Sort by month name. Recognizes three-letter abbreviations (Jan, Feb, ..., Dec) and full month names (January, February) if your system locale supports it. Perfect for date columns that use month names instead of numbers.

• h: Sort human-readable numbers (like 1K, 2M, 3G for kilobytes/megabytes/gigabytes). It parses the suffixes and sorts values numerically, so 100K comes before 2M.

• n: Use numeric value sorting (integer-focused). Treats the key as an integer, so 10 comes after 2 (unlike default lexicographical sort where 10 would come before 2).

• R: Sort the key randomly (based on the key’s hash). Rows with identical keys stay grouped, but their order relative to other groups is randomized—different from a full shuffle of all rows.

• r: Reverse the sort order for this key. If your global sort is ascending, -k2,2r will make the second key sort in descending order instead.

• V: Version sort. This sorts alphanumeric strings in a version-aware way—for example, v1.9 comes before v1.10, which doesn’t happen with default lexicographical sort.

Quick Example of Combined Usage #

Suppose you have a file data.txt with:

apple  100K  Jan
Banana 2M    Feb
cherry 50K   Dec

Running this command:

sort -k2,2hr -k3,3M data.txt

Will:

1. Sort the second key (size column) in human-readable reverse order (so 2M comes first, then 100K, then 50K)
2. For rows with the same size, sort by the third key (month) in month order (Jan before Feb before Dec)

内容的提问来源于stack exchange，提问作者user276072