HBase data model (column families) in Hadoop - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Column Families in HBase

In HBase, what is the primary purpose of grouping columns into column families?

A To physically store related columns together on disk for efficient access

B To enforce data type constraints on columns within the family

C To automatically replicate data across different clusters

D To define user access permissions for individual columns

💡 Hint

Think about how HBase organizes data on disk to improve read/write performance.

Column families group related columns so that their data is stored together physically on disk. This improves access speed because related data is read or written in blocks.

❓ dax_lod_result

intermediate

Column Family Storage Impact on Data Retrieval

Consider an HBase table with two column families: info and metrics. If you query only columns from the info family, what is the expected impact on data retrieval performance?

A Only the <strong>info</strong> family data blocks are read, improving query speed

B Both <strong>info</strong> and <strong>metrics</strong> families are read, no performance gain

C The entire row is scanned regardless of column family, slowing down the query

D Data from <strong>metrics</strong> family is cached automatically, speeding up retrieval

💡 Hint

Recall how HBase stores column families separately on disk.

HBase stores each column family separately, so querying columns from one family reads only that family's data blocks, improving performance.

❓ data_modeling

advanced

Designing Column Families for Efficient Writes

You are designing an HBase table for a sensor data application. Sensors send frequent updates for temperature and humidity. Which column family design will optimize write performance?

A Create one column family for all sensor data (temperature and humidity together)

B Create separate column families for temperature and humidity data

C Create a column family for metadata and store sensor data as JSON in one column

D Use one column family but store temperature and humidity in different rows

💡 Hint

Think about how HBase handles writes to different column families.

Separate column families allow HBase to write data independently and efficiently, reducing write amplification and improving performance for frequent updates.

🔧 Debug

advanced

Identifying Column Family Misconfiguration

An HBase table has two column families: cf1 and cf2. After adding many columns to cf1, you notice slow read performance. Which misconfiguration is most likely causing this?

A Using different column families for unrelated data

B Column families are not compressed, causing slow reads

C Too many columns in a single column family causing large data blocks

D Row keys are not unique, causing read conflicts

💡 Hint

Consider how column family size affects disk I/O.

Having too many columns in one column family creates large data blocks, which slows down reads because more data than needed is loaded.

🎯 Scenario

expert

Optimizing Column Family Design for Mixed Access Patterns

Your HBase table stores user profiles and their activity logs. Profiles are read frequently but updated rarely. Activity logs are written frequently but read less often. How should you design column families to optimize both read and write performance?

A Store profiles in one table and activity logs in a separate HBase table

B Use one column family for both profiles and activity logs to simplify schema

C Store profiles as JSON in one column family and activity logs as separate columns in another

D Create separate column families: one for profiles and one for activity logs

💡 Hint

Think about how HBase handles read and write workloads per column family.

Separating profiles and activity logs into different column families allows tuning each for their access patterns, improving overall performance.