DVC (Data Version Control) basics in MLOps - Step-by-Step Execution

Process Flow - DVC (Data Version Control) basics

Initialize Git repo

↓

Initialize DVC in repo

↓

Add data files to DVC tracking

↓

Commit changes to Git

↓

Push data and code to remote storage

↓

Pull data and code from remote when needed

This flow shows how you start with a Git repo, add DVC to track data files, commit changes, and push or pull data versions.

Execution Sample

MLOps

git init

dvc init

dvc add data.csv

git add .
git commit -m "Add data with DVC"
dvc push

This code initializes Git and DVC, tracks a data file with DVC, commits changes, and pushes data to remote storage.

Process Table

Step	Command	Action	Result
1	git init	Create new Git repository	Empty Git repo created
2	dvc init	Initialize DVC in repo	.dvc folder and config created
3	dvc add data.csv	Track data.csv with DVC	data.csv.dvc file created, data tracked
4	git add .	Stage all changes for Git	All changes staged
5	git commit -m "Add data with DVC"	Commit changes to Git	Commit saved with message
6	dvc push	Upload data files to remote storage	Data files uploaded to remote
7		End of process	Data and code versioned and stored

💡 Process stops after data and code are versioned and pushed to remote storage

Status Tracker

Variable	Start	After Step 2	After Step 3	After Step 5	After Step 6
Git repo	None	Initialized	Initialized	Commit with DVC files	Commit with DVC files
DVC config	None	Created	Created	Created	Created
Tracked data	None	None	data.csv tracked	data.csv tracked	data.csv pushed to remote

Key Moments - 3 Insights

Why do we need both Git and DVC commands?

Git tracks code and small files, while DVC tracks large data files separately. Execution table rows 1-2 show Git and DVC initialization, and rows 3-6 show how both work together.

What does 'dvc add' actually do to the data file?

'dvc add' does not change the data file but creates a .dvc file that tracks its version. See execution table row 3 where data.csv.dvc is created.

Why do we run 'git commit' after 'dvc add'?

Because DVC creates tracking files that Git needs to save. Without committing, Git won't remember the data version info. This is shown in rows 4 and 5.

Visual Quiz - 3 Questions

Look at the execution table, what is the result after running 'dvc init'?

A Empty Git repository created

B data.csv.dvc file created

C .dvc folder and config created

D Data files uploaded to remote

💡 Hint

Check row 2 under Result column in the execution table

At which step is the data file actually tracked by DVC?

A Step 1

B Step 3

C Step 2

D Step 6

💡 Hint

Look at the Action column for 'dvc add data.csv' in the execution table

If you skip 'git commit' after 'dvc add', what will happen?

A Git won't save the DVC tracking files

B Data files won't be tracked by DVC

C Data files won't upload to remote

D DVC will fail to initialize

💡 Hint

Refer to key moment about why 'git commit' is needed after 'dvc add'

Concept Snapshot

DVC basics:
- Use 'git init' to start Git repo
- Use 'dvc init' to add DVC
- Use 'dvc add <file>' to track data files
- Commit changes with Git to save DVC tracking
- Use 'dvc push' to upload data to remote
- Use 'dvc pull' to retrieve data versions

Full Transcript

This lesson shows how to use DVC with Git to version control data files. First, you create a Git repository with 'git init'. Then, you add DVC support using 'dvc init'. Next, you track a data file using 'dvc add data.csv', which creates a tracking file but does not change the data itself. After that, you stage and commit all changes with Git commands 'git add .' and 'git commit'. Finally, you upload the data files to remote storage using 'dvc push'. This process helps keep your data and code versions in sync and easy to share or reproduce.