{ "componentChunkName": "component---src-templates-issues-tsx", "path": "/issues/502", "result": {"data":{"issuesJson":{"id":"342436d1-5acb-5406-a50e-f58393d9bf86","title":"chmln/sd: 与sed命令类似，但提供更加直观的查找和替换能力","number":502,"bodyHTML":"
\n



\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
Reposchmln/sd
Views\"views\"
Stars\"stars\"
Forks\"forks\"
License\"license\"
UpdatedAt\"last-commit\"
CreatedAt\"create-at\"
\n\n

\"\"

\n\n
\n

sd - search & displace

\n

sd is an intuitive find & replace CLI.

\n

The Pitch

\n

Why use it over any existing tools?

\n

Painless regular expressions.   sd uses regex syntax that you already know from JavaScript and Python. Forget about dealing with quirks of sed or awk - get productive immediately.

\n

String-literal mode.   Non-regex find & replace. No more backslashes or remembering which characters are special and need to be escaped.

\n

Easy to read, easy to write.   Find & replace expressions are split up, which makes them easy to read and write. No more messing with unclosed and escaped slashes.

\n

Smart, common-sense defaults.   Defaults follow common sense and are tailored for typical daily use.

\n

Comparison to sed

\n

While sed does a whole lot more, sd focuses on doing just one thing and doing it well. Here are some cherry-picked examples where sd shines.

\n

Simpler syntax for replacing all occurrences:

\n\n

Replace newlines with commas:

\n\n

Extracting stuff out of strings containing slashes:

\n\n

In place modification of files:

\n\n

Benchmarks

\n

Simple replacement on ~1.5 gigabytes of JSON

\n
hyperfine --warmup 3 --export-markdown out.md \\\n  'sed -E \"s/\\\"/'\"'\"'/g\" *.json > /dev/null' \\\n  'sed    \"s/\\\"/'\"'\"'/g\" *.json > /dev/null' \\\n  'sd     \"\\\"\" \"'\"'\"'\"   *.json > /dev/null'
\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CommandMean [s]Min…Max [s]
sed -E \"s/\\\"/'/g\" *.json > /dev/null2.338 ± 0.0082.332…2.358
sed \"s/\\\"/'/g\" *.json > /dev/null2.365 ± 0.0092.351…2.378
sd \"\\\"\" \"'\" *.json > /dev/null0.997 ± 0.0060.987…1.007
\n

Result: ~2.35 times faster

\n

Regex replacement on a ~55M json file:

\n
hyperfine --warmup 3 --export-markdown out.md \\\n  'sed -E \"s:(\\w+):\\1\\1:g\"    dump.json > /dev/null' \\\n  'sed    \"s:\\(\\w\\+\\):\\1\\1:g\" dump.json > /dev/null' \\\n  'sd     \"(\\w+)\" \"$1$1\"      dump.json > /dev/null'
\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
CommandMean [s]Min…Max [s]
sed -E \"s:(\\w+):\\1\\1:g\" dump.json > /dev/null11.315 ± 0.21511.102…11.725
sed \"s:\\(\\w\\+\\):\\1\\1:g\" dump.json > /dev/null11.239 ± 0.20811.057…11.762
sd \"(\\w+)\" \"$1$1\" dump.json > /dev/null0.942 ± 0.0040.936…0.951
\n

Result: ~11.93 times faster

\n

Installation

\n

Install through
\ncargo with
\ncargo install sd, or through various package managers

\n

\"Packaging

\n

Quick Guide

\n
    \n
  1. \n

    String-literal mode. By default, expressions are treated as regex. Use -F or --fixed-strings to disable regex.

    \n
    > echo 'lots((([]))) of special chars' | sd -F '((([])))' ''\nlots of special chars
    \n
    2. \n
  2. \n

    Basic regex use - let's trim some trailing whitespace

    \n
    > echo 'lorem ipsum 23   ' | sd '\\s+$' ''\nlorem ipsum 23
    \n
    3. \n
  3. \n

    Capture groups

    \n

    Indexed capture groups:

    \n
    > echo 'cargo +nightly watch' | sd '(\\w+)\\s+\\+(\\w+)\\s+(\\w+)' 'cmd: $1, channel: $2, subcmd: $3'\ncmd: cargo, channel: nightly, subcmd: watch
    \n

    Named capture groups:

    \n
    > echo \"123.45\" | sd '(?P<dollars>\\d+)\\.(?P<cents>\\d+)' '$dollars dollars and $cents cents'\n123 dollars and 45 cents
    \n

    In the unlikely case you stumble upon ambiguities, resolve them by using ${var} instead of $var. Here's an example:

    \n
    > echo '123.45' | sd '(?P<dollars>\\d+)\\.(?P<cents>\\d+)' '$dollars_dollars and $cents_cents'\n and\n\n> echo '123.45' | sd '(?P<dollars>\\d+)\\.(?P<cents>\\d+)' '${dollars}_dollars and ${cents}_cents'\n123_dollars and 45_cents
    \n
    4. \n
  4. \n

    Find & replace in a file

    \n
    > sd 'window.fetch' 'fetch' http.js
    \n

    That's it. The file is modified in-place.

    \n

    To preview changes:

    \n
    > sd -p 'window.fetch' 'fetch' http.js
    \n
    5. \n
  5. \n

    Find & replace across project

    \n

    This example uses fd.

    \n

    Good ol' unix philosophy to the rescue.

    \n
    fd --type file --exec sd 'from \"react\"' 'from \"preact\"'
    \n

    Same, but with backups (consider version control).

    \n
    fd --type file --exec cp {} {}.bk \\; --exec sd 'from \"react\"' 'from \"preact\"'
    \n
    6. \n
\n

Edge cases

\n

sd will interpret every argument starting with - as a (potentially unknown) flag.
\nThe common convention of using -- to signal the end of flags is respected:

\n
$ echo \"./hello foo\" | sd \"foo\" \"-w\"\nerror: Found argument '-w' which wasn't expected, or isn't valid in this context\n\nUSAGE:\n    sd [OPTIONS] <find> <replace-with> [files]...\n\nFor more information try --help\n$ echo \"./hello foo\" | sd \"foo\" -- \"-w\"\n./hello -w\n$ echo \"./hello --foo\" | sd -- \"--foo\" \"-w\"\n./hello -w
\n

Escaping special characters

\n

To escape the $ character, use $$:

\n
❯ echo \"foo\" | sd 'foo' '$$bar'\n$bar
","updatedAt":"2025-05-22T01:47:38Z","upvoteCount":null,"author":{"login":"eryajf","avatarUrl":"https://avatars.githubusercontent.com/u/33259379?u=e4a4090a38ac2473aaed4ef9945233636776c6c3&v=4","url":"https://github.com/eryajf"},"category":null,"labels":{"edges":[{"node":{"name":"更多","color":"25B472"}},{"node":{"name":"命令行工具","color":"e05879"}},{"node":{"name":"chmln","color":"645bde"}}]},"comments":{"edges":[]}}},"pageContext":{"number":502,"previous":{"title":"sdkman/sdkman-cli: 在任何基于 Unix 的系统上管理多个软件开发工具包的多个版本的工具","number":501},"next":{"title":"bitnami-labs/sealed-secrets: 一款 Kubernetes 控制器並用于实现单向加密 Secrets 的工具","number":503}}}, "staticQueryHashes": ["151096407","2861350382"]}