ssed v1.5.0
Usage
> ssed --help
> something | ssed [commands]
> <input ssed --explain [commands]
> ssed [commands] <fileAll commands use a "g/re/p"-like syntax:
cmd/match[/replace][/flags]About
ssed is an alternative to 'sed'. Not a drop-in replacement, but used for similar tasks. I wanted more familiar and modern Regex support, and have kept adding more functions as ssed became more and more my go-to text manipulation tool.
Line rules transform on a line-by-line basis and, if only line rules are employed, support STDIN streaming.
Document rules transform the entire document. STDIN is read to completion before document rules are executed.
Like sed, and symbol can be used to separate the rule arguments, but unlike sed some separators change the behaviour of the command. The special separators *are ':' and '`'.
Separators
The separator can change the behaviour of the line rule. Rules that support pattern matches likely also support line and literal matches.
/ Default separator, indicates Regex. Supports flags.
: Indicates a line-number rule
\` Indicates a literal string match
All other delimiters indicate Regex
s/\\w+/bar regex match
s|\\w+|bar alternative separator
s|\\w+|bar|i with case insensitive flag enabled
s\`foo\`bar literal match against 'foo'
s:1:bar matches line 1
s{\\w+}bar brackets can also be used (regex match)
s{\\w+}bar}i but they don't have to matchOptions
--explain Explains what the command will do.
--diff Only show differences
--no-diff Do not show differences
--color Enable ANSI colors (true if stdout is a TTY)
--no-color Disable ANSI colors
--ls Read file names from standard input (\`ls | ssed --ls\`)
--input=fromfile Use 'fromfile' as input.
--input fromfile
--input=fromfile1,fromfile2 Run against multiple files.
--input fromfile1 --input fromfile2
--no-input Do not use any file as input.
--write Write each changed file in place.
--write=tofile Provide a destination file name.
--write-rename=%.backup Use the input filename in place of '%'
--no-write Do not write each changed file in place.
--interactive Ask before writing the file
--no-interactive Do not ask before writing the file
--dry-run (-n) Show which files would be affected.
--no-dry-run Do not show which files would be affected.Line Rules
Line rules operate on every line. Commands 'on', 'after' and 'off' share the on/off state.
s/$search/$replace Replace the first instance of 'search' with 'replace'
(aka sub/…/…)
g/$search/$replace Replace every instance of 'search' with 'replace'
(aka gsub/…/…)
t/$pattern (aka take/…) Only print the matching part of the line, or print the entire line if 'pattern' doesn't match
r/$pattern (aka rm/…) Remove the matching part of the line, or print the entire line if 'pattern' doesn't match
1/$pattern Only print the first group of the match
1 Only print the first "column" (columns are separated by whitespace)
cols/$pattern/$columns Split the line by 'pattern' (default is /\\s+/) and print columns (joined by ' ')
cols/$pattern/$columns/$joiner Same, but columns are joiner by $joiner
on/$pattern Start printing on the line where 'pattern' is matched
off/$pattern Stop printing on the line where 'pattern' is matched
after/$pattern Start printing on the line *after* 'pattern' is matched
toggle/$pattern Turn printing off at the matching line, then off, then on...
p/$pattern (aka print/…) Only print lines that match 'pattern'
d/$pattern (aka del/…) Do not print lines that match 'pattern'
!p/$pattern (aka !print/…) Alias for d/del because I can't remember one (d) and always remember the other (p)
prepend/$text (aka prefix/…)
append/$text (aka suffix/…) Adds text to the beginning (prepend) or end (append) of the line
surround/$prefix/$suffix
uniq (aka unique) Only print unique lines
uniq/$pattern Only print matching lines, and only unique matches (uniqueness is determined by the matching regex)Document Rules
sublines/$pattern/$replace For every line that matches, insert one line from replace. Remaining lines will
(aka sl/…) be inserted into the last matched line. Does not do regex replacement.
sort sort/$pattern Sort the lines alphabetically using localeCompare.
If a pattern is provided, the matching part of the line will be used, but the
entire line will be printed.
sortn sortn/$pattern Sort the lines numerically. If no pattern is given, it matches the *first* number.
reverse Obvious
line Prepend each line with the line number
begin:$prepend
border:$prepend:$append Prepend, append, or surround the document (add header/footer)
end:$append
join join/$separator Join lines with a space (or optional separator)Conditions
You can apply rules only under certain conditions, e.g. 'if/{pattern} {rule}' only runs `rule` only lines that match `pattern`.
You can even group rules using `{ rules… }`, and rules can be negated with a preceding '!'.
if/$pattern $rule
!if/$pattern $rule
between/$onPattern/$offPattern $ruleExample: ssed 'if/first-name|last-name' { s/colin/REDACTED/i s/gray/REDACTED/i }
Line Number rules
Using the special delimiter ':' you can apply most rules on line numbers instead of line content. In the case of the 'sub' command, the entire line will be replaced with the literal text.
For example s:1:replace Replaces line 1 with the word "replace" p:1 Only print line 1
Not all rules support this feature, but typically any rule that could support it, does
Line numbers can be expressed as a single number, a range, an open range, a modulo operation (with offset), and a comma-separated list of line rules. p:1 Only matches the line number (only matches line 1) p:%2 Matches lines that are modulo-N (even lines) p:%2-1 Matches lines that are modulo-N minus Y (odd lines) p:1,3,5 Matches the listed line numbers (and only these) p:1-5 Matches the range of number, inclusive (1,2,3,4,5) p:9- Matches the line number and all subsequent lines (lines 9 and onward) p:-9 Matches lines up to and including the line number (lines 1-9) p:1-5,10-15,20,30+ Line rules can be mixed and matched