DEV Community: Sharma Nitesh

Python Database Migrations with Alembic — Complete Guide (2026)

Sharma Nitesh — Thu, 30 Jul 2026 03:18:34 +0000

If you're building a Python app with a database,
you need migrations.

Migrations let you change your database schema
without losing data. Add a column, rename a table,
drop an index — all safely and reversibly.

Alembic is the standard migration tool for
Python + SQLAlchemy apps.

This guide covers everything you need to know.

What is Alembic?

Alembic is a database migration tool for Python.
It works with SQLAlchemy and supports:

PostgreSQL
MySQL
SQLite
Oracle
Microsoft SQL Server

Think of it like Git — but for your database schema.
Every change is versioned, reversible, and trackable.

Install Alembic

pip install alembic sqlalchemy

Verify installation:
alembic --version

Project Setup

Create a new project:
alembic init migrations

This creates:
migrations/
env.py
script.py.mako
versions/
alembic.ini

Configure Database URL

Open alembic.ini and set your database URL:

For SQLite (easiest to start):
sqlalchemy.url = sqlite:///myapp.db

For PostgreSQL:
sqlalchemy.url = postgresql://user:password@localhost/dbname

For MySQL:
sqlalchemy.url = mysql+pymysql://user:password@localhost/dbname

Define Your Models

Create models.py:

from sqlalchemy import Column, Integer, String,
DateTime, create_engine
from sqlalchemy.orm import DeclarativeBase
from datetime import datetime

class Base(DeclarativeBase):
pass

class User(Base):
tablename = "users"

id = Column(Integer, primary_key=True)

email = Column(String(255), unique=True)

name = Column(String(100))

created_at = Column(DateTime, default=datetime.utcnow)

Configure env.py

Open migrations/env.py and add your models:

from models import Base

target_metadata = Base.metadata

Create Your First Migration

Auto-generate migration from your models:
alembic revision --autogenerate -m "create users table"

This creates a file in migrations/versions/ like:
2026_07_27_abc123_create_users_table.py

It looks like this:

def upgrade() -> None:
op.create_table('users',
sa.Column('id', sa.Integer(), nullable=False),
sa.Column('email', sa.String(255), nullable=True),
sa.Column('name', sa.String(100), nullable=True),
sa.Column('created_at', sa.DateTime(), nullable=True),
sa.PrimaryKeyConstraint('id'),
sa.UniqueConstraint('email')
)

def downgrade() -> None:
op.drop_table('users')

Run Migration (Upgrade)

Apply migration to database:
alembic upgrade head

Check current version:
alembic current

View migration history:
alembic history

Add a Column

Add phone field to User model:
phone = Column(String(20), nullable=True)

Generate migration:
alembic revision --autogenerate -m "add phone to users"

Run it:
alembic upgrade head

Rollback Migration (Downgrade)

Undo last migration:
alembic downgrade -1

Undo all migrations:
alembic downgrade base

Go to specific version:
alembic downgrade abc123

Common Alembic Commands

alembic upgrade head # Apply all migrations
alembic downgrade -1 # Undo last migration
alembic current # Show current version
alembic history # Show all migrations
alembic heads # Show latest versions
alembic revision --autogenerate -m "message" # Create migration

Best Practices

Always review auto-generated migrations
before running them
Never edit a migration after it's been
run in production
Keep migrations small and focused —
one change per migration
Always test downgrade() works before
deploying to production
Commit migration files to git — they
are part of your codebase

Common Errors and Fixes

"Target database is not up to date":
alembic upgrade head

"Can't locate revision":
alembic history --verbose

"No changes detected" on autogenerate:
Check env.py imports your models correctly

Practice Alembic in Your Browser

Want to practice SQLAlchemy and database
concepts without setting up a local environment?

Try pyrun.in — SQLite playground and Python
terminal run directly in your browser.

No install. No setup. Just open and code.

Free to start → pyrun.in

Built pyrun.in — browser-based Python learning
in Mumbai. Questions welcome in comments.

Learn Python in Hindi — Complete Beginner Guide (2026)

Sharma Nitesh — Sat, 25 Jul 2026 11:14:58 +0000

पायथन सीखना चाहते हो?

Most Python tutorials are in English.
But millions of Indian learners think in Hindi.

That's the gap PyRun is fixing.

This guide is for Hindi-speaking beginners
who want to learn Python — in the language
they actually think in.

Python क्यों सीखें? (Why Learn Python?)

Python is the most beginner-friendly
programming language in 2026.

इसके फायदे:

Simple English-like syntax
Used in AI, data science, automation
Highest paying tech skills in India
No prior coding experience needed

Salary after learning Python:

Data Analyst: ₹4-8 LPA
Python Developer: ₹5-12 LPA
ML Engineer: ₹8-20 LPA

सबसे बड़ी समस्या (The Biggest Problem)

Python install करना beginners के लिए
सबसे बड़ी रुकावट है।

"PATH not found"
"pip is not recognized"
"which version do I download?"

ये errors देखकर ज़्यादातर beginners
lesson 1 से पहले ही छोड़ देते हैं।

Solution — Browser में Python चलाओ

pyrun.in खोलो।

Python आपके browser में already चल रहा है।
कुछ install नहीं करना। कोई setup नहीं।
बस tab खोलो और code लिखना शुरू करो।

पहले 5 lessons बिल्कुल free हैं।
Hindi में interface switch करने का option है।

Python Basics — हिंदी में

Hello World

print("नमस्ते दुनिया!")
print("Hello, World!")

Variables (चर)

naam = "Rahul"
umar = 20
sheher = "Mumbai"

print(naam)
print(umar)
print(sheher)

Numbers (संख्याएं)

a = 10
b = 5

print(a + b) # जोड़ = 15
print(a - b) # घटाव = 5
print(a * b) # गुणा = 50
print(a / b) # भाग = 2.0

If-Else (अगर-तो)

marks = 75

if marks >= 60:
print("Pass हो गए!")
else:
print("फिर से कोशिश करो")

Loop (चक्र)

for i in range(1, 6):
print(f"Number: {i}")

Output:

Number: 1

Number: 2

Number: 3

Number: 4

Number: 5

Function (कार्य)

def greet(naam):
return f"नमस्ते, {naam}!"

print(greet("Priya"))

Output: नमस्ते, Priya!

PyRun में कैसे Practice करें

pyrun.in खोलो
ऊपर language में Hindi select करो
कोई भी lesson choose करो
Code editor में type करो
Run button दबाओ
Output देखो

बस इतना ही। कोई installation नहीं।
कोई PATH error नहीं।

30 दिनों में Python सीखने का Plan

Week 1 — Basics:

Variables और data types
Print statements
Basic math operations

Week 2 — Control Flow:

If-else conditions
For and while loops
Functions

Week 3 — Data Structures:

Lists (सूचियां)
Dictionaries (शब्दकोश)
Tuples और Sets

Week 4 — Projects:

Calculator बनाओ
To-do list app
Data analysis script

Free Resources

pyrun.in — 93 lessons, Hindi support,
browser-based, no install

आज ही शुरू करो

Python सीखने के लिए कोई special computer
नहीं चाहिए।
कोई paid course नहीं चाहिए।
कोई installation नहीं चाहिए।

बस pyrun.in खोलो और आज ही
अपना पहला Python program लिखो।

Free है। Hindi में है। Browser में है।

pyrun.in 🐍

Solo founder from Mumbai.
Built pyrun.in in 90 days for Indian learners.
Comments and feedback welcome!

Python vs JavaScript in 2026 — Which Should You Learn First?

Sharma Nitesh — Wed, 22 Jul 2026 15:54:56 +0000

Every beginner asks the same question.

"Should I learn Python or JavaScript first?"

In 2026, the answer is clearer than ever.
But it depends on what you want to build.

Let's break it down honestly.

The One-Line Summary

Python → Data, AI, automation, backend
JavaScript → Web, frontend, fullstack, apps

If you don't know what you want to build yet —
learn Python first. Here's why.

Syntax: Which is Easier?

Python wins. No contest.

Python:


python
name = "Rahul"
print(f"Hello, {name}!")
![ ](https://dev-to-uploads.s3.us-east-2.amazonaws.com/uploads/articles/zccoqco7yrhw2w1glnsa.png)

How to Use Python Without Installing Anything (2026)

Sharma Nitesh — Sun, 19 Jul 2026 09:38:25 +0000

If you've ever tried to learn Python, you know
the drill.

Go to python.org. Download the installer.
Run it. Hope nothing breaks.

Then comes the moment every beginner dreads:

Learn Python Without Installing Anything — In-Browser Python in 2026

Sharma Nitesh — Sun, 12 Jul 2026 10:06:00 +0000

For the last 15 years, the answer to "how do I start learning Python?" was: install Python, install an editor, install pip packages, learn what a venv is, fix your PATH, and then write your first print("hello"). That is 3 hours of setup before you type a single character of actual code. Half of beginners quit during the install step.

In 2026, that friction is finally gone. You can now run real Python — not a toy, not a subset, actual CPython — inside a browser tab. No install, no account required for basics, no cloud VM billing you when you forget to turn it off. It works on a Chromebook. It works on a 5-year-old Android tablet. It works on the library computer where you do not have admin rights.

What is Pyodide and why it changed everything

Pyodide is CPython compiled to WebAssembly. The Mozilla research team started it in 2018, and by 2024 it was mature enough that Jupyter itself started shipping a browser version (JupyterLite) built on top. In 2026, the version most in-browser platforms run is Pyodide 0.28, which ships Python 3.13 with numpy, pandas, scipy, matplotlib, scikit-learn, requests, beautifulsoup, and roughly 200 other packages precompiled and ready to import.

The technical trick: WebAssembly is a low-level binary format that browsers execute at near-native speed. Pyodide compiles the C source of CPython into a .wasm binary — around 12 MB gzipped — that the browser downloads once, caches, and then runs a full Python interpreter inside a sandbox on your machine. Not a server. Not a cloud VM. Your laptop, your CPU.

Is browser Python actually fast enough in 2026?

Short answer: for learning and for 90% of scripting, yes.

Task	Native Python	Pyodide (browser)	Ratio
Fibonacci(30) recursive	0.18s	0.42s	2.3× slower
Sort 1M integers	0.31s	0.68s	2.2× slower
pandas: groupby on 100k rows	0.09s	0.14s	1.5× slower
numpy matrix multiply 1000×1000	0.04s	0.05s	1.25× slower
requests.get() to a public API	0.28s	0.31s	Roughly equal

The numpy and pandas gap is tiny because those libraries are C extensions and Pyodide runs them as compiled WebAssembly modules, not through the Python interpreter.

Five real use cases where you never need to install Python locally

1. Learning Python from scratch

Every friction point in the traditional install-first flow is removed. Beginners write, run, and iterate — the loop that actually builds skill.

temperatures = [28, 31, 33, 29, 27, 34, 30]
avg = sum(temperatures) / len(temperatures)
hot_days = [t for t in temperatures if t > 30]

print(f"Weekly average: {avg:.1f}C")
print(f"Hot days: {len(hot_days)}")

Output appears in the same tab. No context switching between terminal, editor, and browser.

2. Quick scripts and one-off data munging

Someone sends you a CSV and asks "what is the median?" You do not open a terminal, activate a venv, install pandas, and write a script. You open a tab.

3. Education at scale

Colleges and coaching institutes have spent 20 years running Python labs where 40% of the class period is IT support fixing broken installs. In-browser Python removes that entirely.

4. Sharing runnable snippets

Embed a Python playground in a blog post, a doc, a bug report. The reader runs your code without leaving the page.

5. Mobile and Chromebook development

A huge chunk of first-time coders in India start on Android tablets or Chromebooks. You cannot install CPython on either without jailbreaking or paying for a cloud IDE. In-browser Python is the only real option, and it now works well enough to complete an entire fundamentals course before you ever need a laptop.

The real limits — where browser Python cannot go (yet)

No system access — you cannot open arbitrary files on your disk, spawn subprocesses, or use os.system(...).
CORS-restricted network — requests.get() to a random site often fails unless that site sends the right CORS headers.
Long-running jobs — a 6-hour ML training run is not what this is for.
Native C extensions not precompiled — anything outside the ~200 Pyodide-supported packages requires a workaround.

If you are running production ML pipelines, deploying a Django site, or building a desktop application, you still need a local install. For everything else — learning, scripting, teaching, sharing, mobile — the browser is now the better default.

Try it right now

Full disclosure: I'm the founder of https://pyrun.in, a Python learning platform built on Pyodide 0.28 running Python 3.13. Every lesson runs in the browser. Every project stub runs in the browser. The fundamentals track takes you from print("hello") to writing a real scraper without ever touching a terminal.

You can also drop in and just try Python without signing up: https://pyrun.in.

The takeaway

In-browser Python is no longer a toy. It is the fastest way from "I want to learn Python" to "I wrote a working program" in 2026 — the install-first workflow is now the slower path, not the default one.

If you teach Python, mentor students, or run a coding club — try opening https://pyrun.in in front of them once. The look on someone's face when they realize they just ran real CPython in Safari on their phone is the reason I built this in the first place.

Questions about Pyodide, browser-based Python, or what breaks at scale? Drop them in the comments — I'll answer every one.