Terminal Scenarios

--Scenarios--
Examples, Scenarios and Tricks:

The nothing trick. Creates a file or wipes if it exists
$ > file.txt


The Null Trick
$ : > file.txt
Wipes existing file, or creates it. ":" is a shell reserved character for null
Then open it with default editor app (TextEdit). Dont open .sh scripts
$ open a.txt
.
Old Echo Trick:
$ echo > file.txt    # or
$ echo " " > file.txt
[from UbuntuForums]

Append. With no arguments, echo adds a new line to the end of the file
$ echo >> file.txt
Add 3 new lines to the end:
$ for i in {1..3}; do echo >> file.txt; done
Or
$ sed '$G;$G;$G'

List unhidden folders only, because folders technically end with / :
$ ls -d */
List folders with full path:
$ ls -d ~/*/
List root:
$ ls -d /*/
List folders inclu hidden:
$ find . -maxdepth 1 -type d
List folders and sub-folders:
$ find . -type d

Show every path in $PATH
$ for i in {1..7}; do echo $PATH | cut -d: -f$i; done

Show Hex's IP
$ ifconfig en1 inet
$ ifconfig | grep 'inet '

Save a list of all the filenames in the current folder
$ ls > file.txt
List all files in all folders within the current folder
$ ls *
Display the 10 newest files
$ ls -lt | head

Show (concatenate) the contents of every file in the folder, with newlines:
$ for i in *; do cat $i && echo; done

[for naming. And generating test files? Clean these up]
Brace expansions
$ echo {1..5}
> 1 2 3 4 5
$ echo No{1..5}
> No1 No2 No3 No4 No5
$ echo {Z..A}
> Z Y X W V U T S R Q P O N M L K J I H G F E D C B A
$ echo {a..Z}
>
Make 9 folders named "1" to "9"
$ mkdir {1..9}
It can take many arguments, space separated
$ echo 0{5..9} {10..12}
> 5 6 7 8 9 10 11 12
Binary Trick
$ echo {0..1}{0..1}{0..1}
> 000 001 010 011 100 101 110 111
Make folders with paths
$ mkdir -p test/{1..10}/{1..10}  [w]
Date Trick. Name folders by date. 0 is to pad 01-09. 12 mo in a yr.
$ mkdir {2007..2009}-0{1..9}  {2007..2009}-{10..12}
My Zelda map trick (unused). Makes a big list of 27 names z00_00 - z09_09
$ echo z0{0..9}_0{0..9} > file.txt
Nested brace expansions:
$ echo a{A{1,2},B{3,4}}b
    aA1b aA2b aB3b aB4b  # comma is just a list?

Change 'seq' delimiter from newline to space
$ seq -s' ' 12
Or pipe to translate. (end with a newline)
$ seq 12 | tr '\n' ' '; echo
> 1 2 3 4 5 6 7 8 9 10 11 12

Sections generator:
$ for i in {1..3}; do echo "Question $i" && echo; done

Generate empty text files in folders for OU course. That's 4 Blocks, 6 Parts, each with up to 20 files:
$ touch script.sh && chmod +x script.sh
```
mkdir -p Block{1..4}/Pt{1..6}
sleep 1
touch Block{1..4}/Pt{1..6}/{0..1}{0..9}.txt
```
Or to only do block by block
` touch Block2/Pt{1..6}/{0..1}{0..9}.txt  `
Delete oo.txt in every folder
$ rm Block{1..4}/Part{1..6}/00.txt
Delete every file in every folder
$ rm Block{1..4}/Part{1..6}/*


To duplicate a file 39 times, and number them "01"-"39". [I need "}{0..9" not "1..9" or it skips 10 (09 11)]
$ for i in {0..3}{0..9}; do cp test.ogg "test$i.ogg"; done

To echo one thing into all files in a folder
$ for i in *; do echo '<!--  -->' >> $i; done

Wipe all files
$ for i in *; do : > $i; done

[Use below output as my standard output?]
To newline an echo pattern (seq only does numbers)
$ for i in {1..9}; do echo "pg"$i; done
> ```
pg1
pg2...
```
To echo specific integers
$ for i in 1 4 5; do echo "pg"$i; done

To delete everything except line 8, inline edit all files ending "txt" (no backups)
for f in *txt;
do ( sed '8!d' "$f" ) > foo && mv foo "$f";
done

[test this?] Move the subdirectory "my_dir" and also all ".bak" files in the current working directory's parent directory to an existing directory named "new_dir"
$ mv my_dir ../*.bak  new_dir

Join all the files in the current folder together
$ cat * > bigcat.txt
Or see the results first
$ cat * | less

In a dir I had 3 text files with 1 line of text. I joined them and double spaced
$ cat * > c.txt
$ sed G c.txt > ch.txt

Number all lines (see AllCmds for minutia)
$ nl -ba -w3 -s' ' file > file_new        # my fave. or
$ cat -n
# or
$ sed = in.txt | sed 'N;s/\n/\t/' > out.txt        # or
$ awk '{print NR  "> " $s}' in.txt > out.txt


Four ways to handle file names with spaces in them. From Macworld
1. Drag the directory onto the terminal
2. Use single quotes (or double):
$ cd '/Users/directory with spaces'
3. Prefix the 'space character' with a backslash (keep tabbing):
$ cd /Users/username/temp/directory\ with\ spaces
4. Replace spaces with a * (or a ?)
And a welcome but over-detailed explanation why these work StackOverflow

To strip Youtube transcripts of timestamps, e.g. Hanselman [lol Ytb can do it now]:
[Updt]
$ tr -d :0-9 | tr -s '\n' ' ' | tr -s ' ' ' ' < in.txt > out.txt
H1 - raw
H2 - colons and numerals removed
H3 - newlines changed to spaces, any duplicates are 'squeezed' to singles. done
$ tr -d :0-9 < H1.txt > H2.txt    # delete numbers
$ tr -s '\n' ' ' < H2.txt > H3.txt    # change newlines to spaces. Squeeze to single spaces
$ tr -s ' ' ' ' < in.txt > out.txt    # squeeze to single spaces (if it didnt work)
If that doesnt work then try different line endings '\r\n'

I had a 'table' (from output of Pi's 'lsblk') but copying from Doc to TextEd turned tabs into 4 spaces [23/7]. Term mostly fixed it with:
$ tr -s ' ' '\t' < in.txt > out.txt
But pasting back into GDoc didn't work, cos of PC line endings?

To extract the tags (single words) from a list of links mixed in, do:
$  sed '/http/d' < in.txt > out.txt
Cos in Firefox's Bookmarks Manager I selected all tags and copied. This pasted the tags AND all links tagged with them into a text doc.


4/3/18 Lynxrot. To re-link the html to css after moving folders around, I wanted to edit the path in multiple files.
[Rewrite this? Clearer folders. "~" delimiters]
Replace text, only on lines matching a pattern, in all html files
5104.html were put in 'htmlandcss'
5104.css were in 'style' beside that folder.
html had 'href="5104.css" ' I wanted to change to 'href="../styles/5104.css" '
so
$ sed -i.bak '/link/s/f="/f="..\/styles\//g' *html
# -i overwrites file; .bak backs up orig; substitute 'f="' (of href); only on lines containing 'link'
# I cudv changed the delimiter to _ or ~ to avoid backslashes
$ rm *.bak
# delete backups, but .bak was in .gitignore anyway.
,
And update index.html hyperlinks from 'href="5104.html" ' to 'href="/COAC/html/5104.css" '
so
$ sed -i.bak '/"[0-9]/s/f="/f="COAC\/htmlandcss\//g' index.html
# only worked as filenames started with a number so if href="5 then find "5. needed this cos the ninja pic URL had 4 numbers in it.
Oops I originally put 'href="/COAC/html/'
To remove a slash and correct the path:
$ sed -i.bak 's/\/C/C/g' index.html
,
Next time maybe "git checkout -- file.html" to revert


How to edit a copy-dump of FreeCodeCamp's syllabus, remove blank lines, remove words:
$ sed '/^$/d' f1.txt > f2.txt
$ sed 's/Complete//; s/Incomplete//' f2.txt > f3.txt
.
[See script]
Or if a syllabus has leading junk, first lesson "Learn HTML" (or "Transcript" on Youtube)
Get line number (or "nl -ba" or "cat -n") and delete from line before
$ sed '/Learn/=' f1.txt
Delete 1st 15 lines
$ sed '1,15d'
.
JS Beta
Which line has only a 1?
$ sed -n '/^1$/='
.
Which line has only a number?
$ sed -n '/^[1-9]$/='     [?]

"m"
[Hang on
Numbering Challenges.
First find what line challenge1 is on:
$ sed -n '/Hello/=' FCCSyllabusHTML.txt
> 18
Delete up to there:
$ sed '1,16d' FCCSyllabusHTML.txt > h1.txt
Find what line the last challenge is on:
$ sed -n '$=' h1.txt
> 59
That's how many files to make:
$ touch ch{0..5}{0..9}.txt


28/3/18  I want to rearrange a tab-separated table of keyboard shortcuts. (And I can't alt+drag TextExit?)
[Delete this spaces bit?]
I copied in 2 halfs in to v01.txt and v02.txt
First append new lines to 1st half
$ echo -e '\n\n\n\n' >> v01.txt      # or a for loop?
$ cat  v01.txt v02.txt  >  v03rd.txt
# join em
$ cut -f1 v03rd.txt > v04.txt
$ cut -f3 v03rd.txt > v05.txt
# extract the 1st+3rd columns (1st had descriptions, 3rd Mac shortcuts, 2nd had Windows unwanted)
$ paste v05.txt v04.txt > v06.txt


To swap a list of pairs so 2nd line is atop 1st.
First delete the first line, and all blank lines (optional)
$ sed '1d; /^$/d'  mraw.txt > m1.txt
delete every 2nd line
$ sed 'n;d'  m1.txt > m2.txt
extract every 2nd line
$ sed -n 'n;p'  m1.txt > m3.txt
merge em
$ paste m3.txt m2.txt > m4.txt
replace tabs (paste) with newlines
$ tr '\t' '\n' < m4.txt > m5.txt


Sort a bibliography alphabetically
$ sort in.txt | cat -s | sed G > out.txt

Converting a comma separated inline list into a newline list:
f1.txt - raw "list, like, this"
$ tr ',' '\n' < f1.txt > f2.txt
f2.txt - replaced commas with newlines, but with spaces at start
$ sed 's/^ //'  f2.txt > f3.txt
f3.txt - no spaces at start. done

Converting a newline list into a comma separated  list (opposite):
f1.txt - raw
$ tr '\n' ',' < f1.txt > f2.txt
f2.txt - replaced newlines with commas. Now to put spaces at start
$ sed 's/,/, /'  f2.txt > f3.txt
.
And the opposite:
$ sed 's/, /€/g' < in.txt > f2.txt
$ tr € '\n' < f2.txt > f3.txt


Create a vertical list of words, stripping out all non-alphanumeric characters e.g. punctuation and whitespace [from tr man page]:
$ tr -cs "[:alpha:]" "\n" < file1 > file2
[Could make practice file by pipe into sort+ uniq?]


Log what ROMs I already transferred to Gato (OS RetroPie on SD RECOVERY) by checking what's still on USB stick SLIM:
setup:
$ cd /Volumes/SLIM/retropie/roms/n64
$ ls > ~/GatoGot01.txt
# creates file. Home dir is an easy to reach path
workflow:
$ cd ../gba
$ ls >> ~/GatoGot01.txt
$ echo "" >> ~/GatoGot01.txt
# append new line for readability


[For AllCmds Sed exmpls?]
A rather fiddly way of formatting a playlist dump how I want:
Chiptune Stream's last.fm playlist has a 5 line pattern - album, artist, song, More, n minutes ago.
Manually delete first lines to get pattern. Delete every 5th line
$ sed 'n;n;n;n;d'  raw.txt > f1.txt
Delete every 4th line.
$ sed 'n;n;n;d'  f1.txt > f2.txt
Now only album > artist > song. Print every 3rd line to extract songs
$ sed -n 'n;n;p' f2.txt > f3.txt
Manually add 2 lines at start. Then print every 3rd line to extract artist
$ sed -n 'n;n;p' f2.txt > f4.txt
Make the bridging "by" file
$ for i in {1..20}; do echo 'by' >> f5.txt; done
Paste
$ paste f3.txt f5.txt f4.txt >  f6.txt
Tabs to spaces, newlines to commas, commas to comma spaces
$ cat f6.txt | tr '\t' ' ' | tr '\n' ',' | sed 's/,/, /g'  > f7.txt

[For AllCmds Sed exmpls?]
I copied folder names from a 'parent directory' view of a website. Want to delete trailing slash and space, add a leading space (cos folder tree in a text doc)
Raw: "
folder1/
folder2/
"
$ sed 's`/ ``; s/^/ /'  in.txt > out.txt


Show the total number of files in the current directory and subdirectories. [from?][works but -print not needed. wc shows line count default anyway]
$ find . -type f -print | wc -l
This shows all files (inclu hidden?). word/line/chara count
$ find . | wc









[for Tests?]
Can I use this command?
A guide mentioned a command "dscacheutil". To check if Hex knows it I started typing and it tab-completed.
Or I opened a fresh shell Cmd+n, hit tab trick to see all cmds and it appeared in the list. Q to quit list or close window.









Info Tests [some obsolete]:
$ ls -al  - list all hidden files, with more info
$ <command> --version    # check the version of a CLI app or language e.g. vim, python
$ man <cmd>    - show the manual page for the command
$ type <cmd>    - check the type of a command [lxcmd.org]
$ which <lang>    - shows which sys folder the language is? works for progs
$ echo $PATH - where Hex looks when launching a CLI prog after running a command
$ printenv | less    # shows all environment variables instead of echoing
$ echo $PS1  - show how the prompt is formatted
$ <cmd> <file> | less    - see results of a command
$ apropos <cmd>    # what other cmds have similar functions?
$ ifconfig en1 inet - current IP address and connection
$ ipconfig getifaddr en1
$ ping bbc.co.uk <n>    - Then ctrl+c to stop. Or quantity 10 (lol common pitfall?)(builtin)