Why NoSQL on GCP matters - Why It Works

Introduction

Sometimes data doesn't fit neatly into tables with rows and columns. NoSQL databases let you store and access data in flexible ways. Google Cloud Platform offers NoSQL services that help apps handle big, changing, or complex data easily.

When your app needs to store lots of data that changes quickly, like user profiles or game scores.

When you want to build apps that handle different types of data, like text, images, or location info, without strict rules.

When you need your app to respond fast to many users at the same time, like chat apps or online shopping carts.

When you want to avoid managing database servers and let Google handle scaling and backups.

When your data structure might change often and you want to avoid costly database redesigns.

Commands

This command creates a Firestore NoSQL database in the us-central1 region on Google Cloud. Firestore is a flexible NoSQL database service.

Terminal

gcloud firestore databases create --region=us-central1

Expected Output

Created Firestore database in region us-central1.

→

--region - Specifies the geographic location for the database to reduce latency.

This command creates a composite index on the 'users' collection to speed up queries filtering by last name ascending and age descending.

Terminal

gcloud firestore indexes composite create --collection-group=users --field-config=fieldPath=lastName,order=ASCENDING --field-config=fieldPath=age,order=DESCENDING

Expected Output

Created composite index for collection group users.

→

--collection-group - Specifies the collection group to index.

→

--field-config - Defines the fields and their sort order for the index.

This command adds a new document with user data to the 'users' collection in Firestore.

Terminal

gcloud firestore documents create users/user123 --data='{"firstName":"Alice","lastName":"Smith","age":30}'

Expected Output

Created document users/user123.

This command retrieves the document data for user123 from the 'users' collection to verify it was stored correctly.

Terminal

gcloud firestore documents get users/user123

Expected Output

{

  "firstName": "Alice",

  "lastName": "Smith",

  "age": 30

}

Key Concept

If you remember nothing else, remember: NoSQL on GCP lets you store and access flexible, fast-changing data without managing servers.

Common Mistakes

Trying to create a Firestore database without specifying the region.

The command fails because Firestore requires a region to know where to place your data.

Always include the --region flag with a valid region like us-central1 when creating a Firestore database.

Not creating necessary indexes before running complex queries.

Queries without proper indexes run slowly or fail because Firestore needs indexes to optimize data retrieval.

Create composite or single-field indexes matching your query patterns before querying.

Using SQL thinking when querying NoSQL data.

NoSQL databases like Firestore use document-based queries, not SQL syntax, so SQL commands won't work.

Use Firestore's document and collection commands or SDK methods designed for NoSQL.

Summary

Create a Firestore NoSQL database in a specific region to store flexible data.

Add indexes to speed up queries on your data collections.

Store and retrieve documents easily without managing database servers.

Practice

1. Why is NoSQL often chosen on Google Cloud Platform (GCP) for certain applications?

easy

A. Because it handles flexible and rapidly changing data well

B. Because it requires complex server management

C. Because it only supports structured data

D. Because it is slower than traditional databases

Solution

1. Step 1: Understand NoSQL data handling

   NoSQL databases are designed to store flexible, semi-structured, or unstructured data that can change quickly.

2. Step 2: Relate to GCP use case

   On GCP, NoSQL services like Firestore and Bigtable support fast-changing data without rigid schemas.

3. Final Answer:

   Because it handles flexible and rapidly changing data well -> Option A

4. Quick Check:

   NoSQL = flexible, fast-changing data [OK]

Hint: NoSQL = flexible data, not rigid tables [OK]

Common Mistakes:

• Thinking NoSQL requires complex server management
• Assuming NoSQL only supports structured data
• Believing NoSQL is slower than SQL

2. Which GCP NoSQL service is best described as a document database?

easy

A. Cloud Spanner

B. Cloud Bigtable

C. Cloud SQL

D. Cloud Firestore

Solution

1. Step 1: Identify document database on GCP

   Cloud Firestore stores data as documents in collections, making it a document database.

2. Step 2: Differentiate from other services

   Cloud Bigtable is a wide-column store, Cloud SQL and Spanner are relational databases.

3. Final Answer:

   Cloud Firestore -> Option D

4. Quick Check:

   Document DB on GCP = Firestore [OK]

Hint: Firestore = document DB, Bigtable = wide-column [OK]

Common Mistakes:

• Confusing Bigtable as document DB
• Choosing relational DBs like Cloud SQL
• Mixing Spanner with NoSQL services

3. Given this code snippet using Firestore on GCP:

doc_ref = firestore_client.collection('users').document('user123')
doc_ref.set({'name': 'Alice', 'age': 30})
user = doc_ref.get().to_dict()
print(user['name'])

What will be printed?

medium

A. user123

B. Alice

C. 30

D. None

Solution

1. Step 1: Understand Firestore document creation

   The code sets a document with fields 'name' and 'age' under 'users/user123'.

2. Step 2: Retrieve and print the 'name' field

   The get() call fetches the document, and user['name'] accesses the 'name' value, which is 'Alice'.

3. Final Answer:

   Alice -> Option B

4. Quick Check:

   Print user['name'] = Alice [OK]

Hint: Set then get document fields to print values [OK]

Common Mistakes:

• Printing document ID instead of field
• Confusing field names
• Assuming None if document exists

4. You try to write data to Cloud Bigtable but get an error. Which mistake likely caused it?

table = bigtable_client.table('my-table')
row = table.row('row1')
row.set_cell('cf1', 'name', 'Bob')
row.commit()

medium

A. Missing column family 'cf1' creation before setting cell

B. Using wrong row key format

C. Calling commit() before set_cell()

D. Bigtable does not support string values

Solution

1. Step 1: Check Bigtable column family setup

   Bigtable requires column families to exist before writing cells; 'cf1' must be created first.

2. Step 2: Verify code sequence and data types

   Row key format is valid, commit() is correctly called after set_cell(), and Bigtable supports string values.

3. Final Answer:

   Missing column family 'cf1' creation before setting cell -> Option A

4. Quick Check:

   Column family must exist before writing [OK]

Hint: Create column family before writing cells [OK]

Common Mistakes:

• Forgetting to create column family
• Assuming commit() order is wrong
• Thinking Bigtable rejects strings

5. You want to build a scalable app on GCP that needs to store user profiles with flexible fields and also handle millions of fast writes per second. Which combination of NoSQL services should you choose?

hard

A. Use Cloud SQL for profiles and Cloud Spanner for writes

B. Use Cloud Bigtable for profiles and Cloud Firestore for writes

C. Use Cloud Firestore for flexible profiles and Cloud Bigtable for high-speed writes

D. Use Cloud Spanner for both profiles and writes

Solution

1. Step 1: Match data needs to NoSQL services

   Cloud Firestore handles flexible, document-based user profiles well.

2. Step 2: Match write speed needs

   Cloud Bigtable supports massive scale and very fast writes, ideal for high throughput.

3. Step 3: Evaluate other options

   Cloud SQL and Spanner are relational and less suited for flexible schema or extreme write speed.

4. Final Answer:

   Use Cloud Firestore for flexible profiles and Cloud Bigtable for high-speed writes -> Option C

5. Quick Check:

   Firestore + Bigtable = flexible + fast writes [OK]

Hint: Firestore for flexible data, Bigtable for speed [OK]

Common Mistakes:

• Choosing relational DBs for flexible schema
• Swapping Firestore and Bigtable roles
• Ignoring write speed requirements