Bash Terminal Cheatsheet for Analytics Engineering

bash_cheatsheet.md

🖥️ Bash Terminal Cheatsheet

Fellow-level reference · Quick lookup, no fluff.

---

📁 File & Directory Navigation

Command	What it does
pwd	Print current working directory
ls -lah	List all files (incl. hidden), human-readable sizes
cd -	Jump back to previous directory
find . -name "*.csv" -type f	Recursively find files by pattern
du -sh ./dir	Disk usage of a directory, human-readable
stat file.txt	File metadata (size, permissions, timestamps)
tree -L 2	Visual directory tree, 2 levels deep
ln -s /path/to/target link_name	Create a symlink

---

🔀 Redirection & Piping

Command	What it does
cmd > file.txt	Redirect stdout to file (overwrites)
cmd >> file.txt	Append stdout to file
cmd 2> err.log	Redirect stderr only to file
cmd 2>&1	Merge stderr into stdout
cmd > file.txt 2>&1	Redirect both stdout + stderr to file
cmd1 | cmd2	Pipe stdout of cmd1 into cmd2
cmd | tee file.txt	Pipe stdout AND write to file simultaneously
cmd < input.txt	Feed file as stdin to command

⚠️ Order matters: > file 2>&1 ✅ vs 2>&1 > file ❌ (stderr stays on terminal)

🔬 Deep dive: 2>&1 explained

Part	Meaning
2	File descriptor for stderr
>	Redirection operator
&1	"The destination of file descriptor stdout"

Anatomy: 2>&1 tells the shell — "send stderr to wherever stdout is currently pointing."

# ✅ Correct — redirect stdout to file first, then attach stderr to it
npm run build > build.log 2>&1

# ✅ Pipe both stdout + stderr into grep
npm run build 2>&1 | grep error

# ❌ Wrong — stderr binds to terminal (old stdout), then stdout moves to file
npm run build 2>&1 > build.log

The & before 1 is critical — without it, the shell interprets >1 as "redirect to a file named 1".

---

⚙️ Process Management

Command	What it does
ps aux	List all running processes
ps aux | grep python	Filter processes by name
top / htop	Live process monitor (htop is prettier)
kill -9 <PID>	Force kill process by PID
pkill -f "script_name"	Kill process by name match
cmd &	Run command in background
jobs	List background jobs in current shell
fg %1	Bring background job 1 to foreground
nohup cmd &	Run in background, survives terminal close
Ctrl+Z → bg	Suspend foreground job, resume in background
lsof -i :8080	See what's using port 8080

---

🔍 Data Wrangling (grep, awk, sed, cut)

grep — search text

Command	What it does
grep "pattern" file.txt	Search for pattern in file
grep -r "pattern" ./dir	Recursive search in directory
grep -i "pattern" file	Case-insensitive search
grep -n "pattern" file	Show line numbers
grep -v "pattern" file	Invert match (exclude pattern)
grep -E "err|warn" file	Extended regex (OR match)
grep -c "pattern" file	Count matching lines

awk — column/field processing

Command	What it does
awk '{print $1}' file	Print first column (space-delimited)
awk -F',' '{print $2}' file.csv	Print second column (comma-delimited)
awk 'NR>1' file	Skip header row
awk '$3 > 100' file	Filter rows where col 3 > 100
awk '{sum+=$2} END {print sum}' file	Sum values in column 2

sed — stream editor / find-replace

Command	What it does
sed 's/old/new/' file	Replace first occurrence per line
sed 's/old/new/g' file	Replace all occurrences per line
sed -i 's/old/new/g' file	In-place edit (modifies file directly)
sed -n '5,10p' file	Print lines 5 to 10
sed '/pattern/d' file	Delete lines matching pattern

cut — slice columns

Command	What it does
cut -d',' -f1,3 file.csv	Extract columns 1 and 3 (comma delimiter)
cut -c1-10 file	Extract characters 1–10 from each line

Bonus combos

# Count unique values in CSV column 2
cut -d',' -f2 data.csv | sort | uniq -c | sort -rn

# Extract error lines and save
grep -i "error" app.log | awk '{print $1, $2, $NF}' > errors.txt

# Replace env var in config, in-place
sed -i 's/DB_HOST=localhost/DB_HOST=prod-db/g' .env

---

🌐 Networking & curl

Command	What it does
curl -X GET "https://api.example.com"	Basic GET request
curl -X POST -H "Content-Type: application/json" -d '{"key":"val"}' URL	POST with JSON body
curl -H "Authorization: Bearer $TOKEN" URL	Authenticated request
curl -o file.json URL	Save response to file
curl -s URL	Silent mode (no progress bar)
curl -v URL	Verbose (shows headers, handshake)
curl -L URL	Follow redirects
curl -w "%{http_code}" -o /dev/null -s URL	Print HTTP status code only
wget -q URL -O output.file	Download file quietly
ping -c 4 google.com	Test connectivity (4 packets)
netstat -tulpn / ss -tulpn	List open ports and listeners

---

🐙 Git CLI

Setup & status

Command	What it does
git status	Working tree status
git log --oneline --graph --all	Compact visual branch history
git diff	Unstaged changes
git diff --staged	Staged changes

Branching

Command	What it does
git checkout -b feature/name	Create + switch to new branch
git branch -d branch_name	Delete local branch (safe)
git branch -D branch_name	Force delete local branch
git push origin --delete branch_name	Delete remote branch
git fetch --prune	Sync remote branches, prune deleted

Staging & committing

Command	What it does
git add -p	Interactively stage hunks
git commit --amend --no-edit	Add to last commit without changing message
git reset HEAD~1	Undo last commit, keep changes staged
git reset --hard HEAD~1	⚠️ Undo last commit, discard changes

Stashing

Command	What it does
git stash push -m "wip: feature"	Stash with label
git stash list	List all stashes
git stash pop	Apply latest stash + drop it
git stash apply stash@{2}	Apply specific stash, keep it

Remote & sync

Command	What it does
git pull --rebase	Pull and rebase local commits on top
git push -u origin branch_name	Push + set upstream tracking
git remote -v	Show remote URLs
git cherry-pick <SHA>	Apply a specific commit to current branch

Inspection & recovery

Command	What it does
git blame file.py	Show who changed each line
git bisect start	Binary search for bug-introducing commit
git reflog	Full history of HEAD movements (lifesaver)
git show <SHA>	Show diff + metadata of a commit

---

Generated for a Fellow-level Analytics & Data Engineering workflow.