Vector database operations (CRUD) in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Vector Database Insert Operation

Which statement best describes what happens during the insert operation in a vector database?

A The vector database adds new vectors along with their metadata to the storage, making them available for future queries.

B The vector database removes vectors that match a given query from the storage permanently.

C The vector database compresses all stored vectors to save space without changing their content.

D The vector database updates the similarity scores of existing vectors without changing the stored vectors themselves.

💡 Hint

Think about what it means to add new data to a database.

Inserting means adding new vectors and their metadata into the database so they can be searched later.

❓ Predict Output

intermediate

Output of Vector Deletion Code

What will be the output of the following code snippet that deletes a vector by ID and then queries the database?

Prompt Engineering / GenAI

vectors = {'v1': [0.1, 0.2], 'v2': [0.4, 0.5]}

# Delete vector with ID 'v1'
del vectors['v1']

# Query remaining keys
print(list(vectors.keys()))

A ['v1', 'v2']

B []

C ['v2']

D KeyError

💡 Hint

Deleting a key removes it from the dictionary.

Deleting 'v1' removes it from the dictionary, so only 'v2' remains.

❓ Model Choice

advanced

Choosing the Best Model for Vector Similarity Search

You want to build a vector database that supports fast similarity search on high-dimensional image embeddings. Which model type is best suited for this task?

A A pre-trained transformer model fine-tuned for text generation

B A vector embedding model trained to produce compact, meaningful image feature vectors

C A dense neural network trained for classification tasks

D A clustering algorithm like K-means used directly as a model

💡 Hint

Think about which model creates vectors that capture image features well.

Embedding models trained on images produce vectors that represent image features well for similarity search.

❓ Hyperparameter

advanced

Effect of Changing Vector Search 'k' Parameter

In a vector database query, the parameter k controls how many nearest neighbors are returned. What happens if you increase k from 5 to 20?

A The query will fail because <code>k</code> must be less than 10.

B The query returns fewer vectors, making the search faster but less comprehensive.

C The query ignores the <code>k</code> value and returns all vectors in the database.

D The query returns more vectors, increasing the chance of finding relevant results but taking more time.

💡 Hint

Think about what increasing the number of neighbors means for results and speed.

Increasing k returns more neighbors, which can improve recall but slows down the query.

🔧 Debug

expert

Debugging Vector Update Operation

Given the following code to update a vector in a vector database, what error will it raise?

Prompt Engineering / GenAI

vectors = {'v1': [0.1, 0.2], 'v2': [0.4, 0.5]}

# Attempt to update vector 'v3'
vectors['v3'][0] = 0.9

print(vectors)

A KeyError because 'v3' does not exist in the dictionary

B IndexError because the vector has no elements

C TypeError because list elements cannot be assigned

D No error; the vector 'v3' is created and updated

💡 Hint

Consider what happens when you try to access a key that is not in a dictionary.

Accessing vectors['v3'] raises a KeyError because 'v3' is not a key in the dictionary.