Docs
  • Now
  • 关于
  • 归档
  • 我的工具

    • Home » 开发文档 » sed cheatsheet

    categories: Handbook tags: Linux Geek Tools Shell

    1437字 3分钟

    #sed

    ##tldr

    sh
    sed -i "[range]s/[patern]/[replacement]/[options]" file_*
sed -i "2,5s|[patern]|[replacement]|[options]" file_*

    ###Search for text

    Search for a string and only print the lines that were matched

    sh
    $ sed -n '/hello/p' file.txt

    Case insensitive search

    sh
    $ sed -n '/hello/Ip' file.txt

    ###Sed Usage

    Syntax

    shell
    $ sed [options] command [input-file]

    With pipeline

    sh
    $ cat report.txt | sed 's/Nick/John/g'
    sh
    $ echo '123abc' | sed 's/[0-9]+//g'

    ###Option Examples

    OptionExampleDescription
    -ised -ibak ’s/On/Off/’ php.iniBackup and modify input file directly
    -Esed -E ’s/[0-9]+//g’ input-fileUse extended regular expressions
    -nsed -n ‘3 p’ config.confSuppress default pattern space printing
    -fsed -f script.sed config.confExecute sed script file
    -esed -e ‘command1’ -e ‘command2’ input-fileExecute multiple sed commands

    ###Multiple commands

    sh
    $ echo "hello world" | sed -e 's/h/H/g' -e 's/w/W/g'
Hello World

    Use -e to execute multiple sed commands

    ###Sed script

    sh
    $ echo 's/h/H/g' >> hello.sed
$ echo 's/w/W/g' >> hello.sed
$ echo "hello world" | sed -f hello.sed
Hello World

    Use -f to execute sed script file

    ###Examples

    sh
    $ sed 's/old/new/g' file.txt
$ sed 's/old/new/g' file.txt > new.txt
$ sed 's/old/new/g' -i file.txt
$ sed 's/old/new/g' -i.backup file.txt

    ###Sed Usage

    Syntax

    shell
    $ sed [options] command [input-file]

    With pipeline

    sh
    $ cat report.txt | sed 's/Nick/John/g'
    sh
    $ echo '123abc' | sed 's/[0-9]+//g'

    ###Option Examples

    OptionExampleDescription
    -ised -ibak ’s/On/Off/’ php.iniBackup and modify input file directly
    -Esed -E ’s/[0-9]+//g’ input-fileUse extended regular expressions
    -nsed -n ‘3 p’ config.confSuppress default pattern space printing
    -fsed -f script.sed config.confExecute sed script file
    -esed -e ‘command1’ -e ‘command2’ input-fileExecute multiple sed commands

    ###Multiple commands

    sh
    $ echo "hello world" | sed -e 's/h/H/g' -e 's/w/W/g'
Hello World

    Use -e to execute multiple sed commands

    ###Sed script

    sh
    $ echo 's/h/H/g' >> hello.sed
$ echo 's/w/W/g' >> hello.sed
$ echo "hello world" | sed -f hello.sed
Hello World

    Use -f to execute sed script file

    ###Examples

    sh
    $ sed 's/old/new/g' file.txt
$ sed 's/old/new/g' file.txt > new.txt
$ sed 's/old/new/g' -i file.txt
$ sed 's/old/new/g' -i.backup file.txt

    ##Sed commands

    ###Commands

    CommandExampleDescription
    psed -n ‘1,4 p’ input.txtPrint lines 1-4
    psed -n -e ‘1,4 p’ -e ‘6,7 p’ input.txtPrint lines 1-4 and 6-7
    dsed ‘1,4 d’ input.txtPrint lines except 1-4
    wsed -n ‘1,4 w output.txt’ input.txtWrite pattern space to file
    ased ‘2 a new-line’ input.txtAppend line after
    ised ‘2 i new-line’ input.txtInsert line before

    ###Space commands

    CommandDescription
    nPrint pattern space, empty pattern space, and read next line
    xSwap pattern space with hold space
    hCopy pattern space to hold space
    HAppend pattern space to hold space
    gCopy hold space to pattern space
    GAppend hold space to pattern space

    ###Flags

    sh
    $ sed 's/old/new/[flags]' [input-file]
    FlagDescription
    gGlobal substitution
    1,2...Substitute the nth occurrence
    pPrint only the substituted line
    wWrite only the substituted line to a file
    IIgnore case while searching
    eSubstitute and execute in the command line

    ###Loops commands

    CommandDescription
    b labelBranch to a label (for looping)
    t labelBranch to a label only on successful substitution
    (for looping)
    :labelLabel for the b and t commands (for looping)
    NAppend next line to pattern space
    PPrint 1st line in multi-line
    DDelete 1st line in multi-line

    ###Misc Flags

    FlagDescription
    / | ^ @ ! #Substitution delimiter can be any character
    &Gets the matched pattern
    ( ) \1 \2 \3Group using ( and ).
    Use \1, \2 in replacement to refer the group

    ##Sed examples

    ###Replacing text

    Replace all occurrences of a string

    sh
    $ sed 's/old/new/g' file.txt

    Replace only the nth occurrence of a string

    sh
    $ sed 's/old/new/2' file.txt

    Replace a string only on the 5th line

    sh
    $ sed '5 s/old/new/' file.txt

    Replace “world” with “universe” but only if the line begins with “hello”

    sh
    $ sed '/hello/s/world/universe/' file.txt

    Remove “" from the end of each line

    sh
    $ sed 's/\\$//' file.txt

    Remove all whitespace from beginning of each line

    sh
    $ sed 's/^\s*//' file.txt

    Remove comments. Even those that are at the end of a line

    sh
    $ sed 's/#.*$//' file.txt

    ###Search for text

    Search for a string and only print the lines that were matched

    sh
    $ sed -n '/hello/p' file.txt

    Case insensitive search

    sh
    $ sed -n '/hello/Ip' file.txt

    Search for a string but only output lines that do not match

    sh
    $ sed -n '/hello/!p' file.txt

    ###Appending lines

    Append line after line 2

    sh
    $ sed '2a Text after line 2' file.txt

    Append line at the end of the file

    sh
    $ sed '$a THE END!' file.txt

    Append line after every 3rd line starting from line 3

    sh
    $ sed '3~3a Some text' file.txt

    ###Numbering

    Number line of a file (simple left alignment)

    sh
    $ sed = file.txt | sed 'N;s/\n/\t/'

    Number line of a file (number on left, right-aligned)

    sh
    $ sed = file.txt | sed 'N; s/^/   /; s/ *\(.\{6,\}\)\n/\1  /'

    Number line of file, but only print numbers if line is not blank

    sh
    $ sed '/./=' file.txt | sed '/./N; s/\n/ /'

    Count lines (emulates “wc -l”)

    sh
    $ sed -n '$='

    ###Prepending lines

    Insert text before line 5

    sh
    $ sed '5i line number five' file.txt

    Insert “Example: " before each line that contains “hello”

    sh
    $ sed '/hello/i Example: ' file.txt

    ###Deleting lines

    Delete line 5-7 in file

    sh
    $ sed '5,7d' file.txt

    Delete every 2nd line starting with line 3

    sh
    $ sed '3~2d' file.txt

    Delete the last line in file

    sh
    $ sed '$d' file.txt

    Delete lines starting with “Hello”

    sh
    $ sed '/^Hello/d' file.txt

    Delete all empty lines

    sh
    $ sed '/^$/d' file.txt

    Delete lines starting with “#”

    sh
    $ sed '/^#/d' file.txt

    ###File spacing

    Double space

    sh
    $ sed G

    Delete all blank lines and double space

    sh
    $ sed '/^$/d;G'

    Triple space a file

    sh
    $ sed 'G;G'

    Undo double-spacing

    sh
    $ sed 'n;d'

    Insert a blank line above line which matches “regex”

    sh
    $ sed '/regex/{x;p;x;}'

    Insert a blank line below line which matches “regex”

    sh
    $ sed '/regex/G'

    Insert a blank line around line which matches “regex”

    sh
    $ sed '/regex/{x;p;x;G;}'

    ##Also see

    ###Replacing text

    Replace all occurrences of a string

    TEXT
    $ sed 's/old/new/g' file.txt

    Replace only the nth occurrence of a string

    TEXT
    $ sed 's/old/new/2' file.txt

    Replace replace a string only on the 5th line

    TEXT
    $ sed '5 s/old/new/' file.txt

    Replace “world” with “universe” but only if the line begins with “hello”

    TEXT
    $ sed '/hello/s/world/universe/' file.txt

    Remove "” from the end of each line

    TEXT
    $ sed 's/\\$//' file.txt

    Remove all whitespace from beginning of each line

    TEXT
    $ sed 's/^\s*//' file.txt

    Remove comments. Even those that are at the end of a line

    TEXT
    $ sed 's/#.*$//' file.txt

    ###Search for text

    Search for a string and only print the lines that were matched

    TEXT
    $ sed -n '/hello/p' file.txt

    Case insensitive search

    TEXT
    $ sed -n '/hello/Ip' file.txt

    Search for a string but only output lines that do not match

    TEXT
    $ sed -n '/hello/!p' file.txt

    ###Appending lines

    Append line after line 2

    TEXT
    $ sed '2a Text after line 2' file.txt

    Append line at the end of the file

    TEXT
    $ sed '$a THE END!' file.txt

    Append line after every 3rd line starting from line 3

    TEXT
    $ sed '3~3a Some text' file.txt

    ###Numbering

    Number line of a file (simple left alignment)

    TEXT
    $ sed = file.txt | sed 'N;s/\n/\t/'

    Number line of a file (number on left, right-aligned)

    TEXT
    $ sed = file.txt | sed 'N; s/^/   /; s/ *\(.\{6,\}\)\n/\1  /'

    Number line of file, but only print numbers if line is not blank

    TEXT
    $ sed '/./=' file.txt | sed '/./N; s/\n/ /'

    Count lines (emulates “wc -l”)

    TEXT
    $ sed -n '$='

    ###Prepending lines

    Insert text before line 5

    TEXT
    $ sed '5i line number five' file.txt

    Insert “Example:” before each line that contains “hello”

    TEXT
    $ sed '/hello/i Example: ' file.txt

    ###Deleting lines

    Delete line 5-7 in file

    TEXT
    $ sed '5,7d' file.txt

    Delete every 2nd line starting with line 3

    TEXT
    $ sed '3~2d' file.txt

    Delete the last line in file

    TEXT
    $ sed '$d' file.txt

    Delete lines starting with “Hello”

    TEXT
    $ sed '/^Hello/d' file.txt

    Delete all empty lines

    TEXT
    $ sed '/^$/d' file.txt

    Delete lines starting with “#”

    TEXT
    $ sed '/^#/d' file.txt

    ###File spacing

    Double space

    TEXT
    $ sed G

    Delete all blank lines and double space

    TEXT
    $ sed '/^$/d;G'

    Triple space a file

    TEXT
    $ sed 'G;G'

    Undo double-spacing

    TEXT
    $ sed 'n;d'

    Insert a blank line above line which matches “regex”

    TEXT
    $ sed '/regex/{x;p;x;}'

    Insert a blank line below line which matches “regex”

    TEXT
    $ sed '/regex/G'

    Insert a blank line around line which matches “regex”

    TEXT
    $ sed '/regex/{x;p;x;G;}'