Basics
man {cmd} |
man page for cmd |
type {cmd} |
executable/shell/alias |
help {cmd} |
if there is no man page |
wc |
word count ( -l
ines -w
ords) |
sort {file} |
alphabetically ( -r
everse -n
umeric -u
nique -k
column) |
ls | sort -nk4 |
sort by file size |
mkdir [-p] |
also make -p
arents if not exists |
rm [-r -d -i] |
( -r
ecursive -d
empty directories -i
prompt before every removal) |
mv {source ...} {destination} |
move file(s) or directories or rename (if single file or folder is specified as source & destination) |
cp {source ...} destination |
copy file(s) |
history | less |
open history in pager |
history | grep {term} |
search for term in history |
cat {file1} {file2} |
concatenate & print contents of file(s) |
less {file} |
page file contents |
head -{n} & tail |
print firts / last n lines of file |
tail -f |
follow (eg. for logs) |
tr {search} {replace} |
replace search with replace in stdout (eg. cat msg | tr s S
to uppercase all s
) ( -d
elete {search} instead of replacing) |
Grep
man grep | grep "count" -i -C5 |
grep in man
for "count" |
-i |
insensitive case |
-w |
word search |
-r |
recursive |
-c |
count number of matches |
-A {n} |
end n lines after match |
-B {n} |
start n lines before match |
-C {n} |
start & end n lines before & after match |
-n |
add line number for each match |
-m {n} |
limit to n matches |
-l |
show filenames |
Search for patterns in each file's contents (supports regex!)
grep "{pattern}" {file}
print each line from {file}
that contains {pattern}
egrep
(or grep -E
) - use regex meta-characters (eg. ?
{}
etc.) for their special meaning and not as a normal character
|
|
Shortcuts
ctrl+l |
clear screen |
ctrl+a / ctrl+e |
move to beginning/end of line |
option+← / option+→ |
move left/right one word |
ctrl+t |
swap char with one before |
option+t |
swap word with one before |
ctrl+u |
kill from cursor to beginning of line |
option+d |
kill word forward |
ctrl+w |
kill word backward |
!! |
last command |
!-{n} |
current command minus n |
ctrl+r |
incremental history search |
Some options on some OSes / terminals need to be set with bindkey, eg. bindkey "^[b" backward-word
Streaming / redirection / piping
> |
redirect output to specific file (overwrite!) |
>> |
append new data to specific file |
< |
pass contents of file to stdin |
2> |
redirect stderr to specific file |
2>> |
append stderr to file |
2>&1 |
redirect both sderr and stdout to same file |
command1 | command2 |
pipe (redirect) stdout of 1 to stdin of 2 |
tee |
read stdin and copy both to stdout and a file. |
cat a.txt b.txt | tee both.txt | wc -l |
see above |
If redirecting both stdout and stderr, out has to be first! Eg. cat a.txt b.txt > both.txt 2> error.txt
less
space or f |
next page |
b |
previous page |
enter or down arrow |
scroll one line |
/ {pattern} |
search for pattern |
q |
quit |
Pager to read manuals, docs, etc.
Permissions
r |
file can be read, dir contents listed |
w |
file can be modified, dir contents modified |
x |
file can be treated as a program to be executed, dir can be cd
d into |
- |
file/dir can't be read/modified/executed (depending on where the dash is) |
chmod {permission} {file} |
change permissions for given file |
chmod g+w |
add write permissions for group |
chmod a-w |
remove write permissions for all |
chmod a=r |
set permissions to read ONLY for all |
chmod 020 |
--- -w- --- |
chmod 700 |
rwx --- --- |
chmod 644 |
rw - r-- r-- |
Modes: user, group, others, all
What: -: remove, +: add, =: only
Which: read, write, xecute
|
|
Expansion
~ |
home dir |
~{someuser} |
home dir of someuser |
* |
zero or more characters |
? |
any single character |
[] |
range of characters eg. [123] [0-9] [A-F] |
[^] |
negate range eg. [^0-9]* -> not starts with a digit |
{} |
generate arbitrary strings, eg. touch page{1,2,3}.txt will generate 3 files |
{...} |
generate between range eg. {1..31} {a..e} {2..10..2} (start,stop,step) |
{x,y{1..5},z} |
will expand to x y1 y2 y3 y4 y5 z
|
$((expression)) |
arithmetic expansion ( +
, -
,
, /
, *
, %
) |
$(command) |
command substitution (eg. echo today is $(date)) |
"double quotes" |
respect spacing & ignore special characters except $, \ and ` (backtick) |
'single quotes' |
suppress all forms of substitution |
Before any command runs, some special commands expand to match given files eg.:
ls -l *.txt
will be expanded to contain all files ending with .txt
mv app?.css Styles/
-> move all files matching pattern to Styles directory
Finding files
locate {searchterm} |
find all files containing searchterm. (-ignore case -l
limit to n options -e
only files that exist now) |
find {dir/} |
list every file and directory inside dir/ |
-type d |
only find directories (or f
iles) |
-name "{pattern}" |
find files with given pattern ( case sensitive!) ( iname
for case insensitive) |
-size +1G |
larger than 1gb (or -50M, 20k -> exactly 20k) |
-user {uname} |
files belonging to uname |
-empty |
empty files/folders |
[acm]min {n} |
accessed / changed / modified n minutes ago (+- for greater/less than) |
[acm]time {n} |
a/c/m n * 24 hours ago |
|
logical operators for more complex needs eg. find -type f -not -name "*.html" |
|
execute cmd for each match, eg.: find ~/Downloads -type f -empty -exec ls -l '{}' ';' |
find -type f -name "*.html -exec cp '{} '{}_COPY' ';' |
find each matching file & create a copy with suffix |
-ok command '{}' ';' |
same as above but ask for confirmation each time ( y
/ n
) |
find -name "*.txt" | xargs ls |
equivalent to the -exec ls '{}' ';' |
|
