What DBMS questions are asked in TCS, Infosys, Wipro fresher interviews in 2026?

TCS, Infosys, Wipro, Cognizant, and Capgemini fresher DBMS interviews in 2026 consistently cover six topic clusters: (1) RDBMS fundamentals - what is DBMS, what is RDBMS, differences (DBMS is the general concept of managing databases; RDBMS is DBMS based on the relational model with tables, rows, columns, and ACID transactions), keys (primary, foreign, unique, candidate, super, composite), relationships (one-to-one, one-to-many, many-to-many); (2) Normalization - 1NF, 2NF, 3NF, BCNF, functional dependencies, anomalies, denormalization; (3) SQL - same as the dedicated SQL interview questions (SELECT, JOINs, GROUP BY, subqueries, window functions); (4) Transactions and ACID - atomicity, consistency, isolation, durability, transaction states, commit/rollback, savepoints; (5) Concurrency control - locking (shared, exclusive, two-phase), deadlocks, isolation levels (read uncommitted, read committed, repeatable read, serializable), phantom reads, dirty reads; (6) Indexing and storage - B-tree, B+ tree, hash index, clustered vs non-clustered, when to use indexes, query optimization. TCS NQT typically has 5-8 DBMS MCQs; Infosys InfyTQ has 8-10 plus 1-2 normalization scenarios; Cognizant GenC adds concurrency and indexing questions.

What is the difference between DBMS and RDBMS?

DBMS (Database Management System) is the general concept of software that manages databases - it can be relational, hierarchical, network, object-oriented, or document-based. Examples: MySQL, PostgreSQL, Oracle, MongoDB (NoSQL but still a DBMS), Microsoft Access. RDBMS (Relational Database Management System) is a specific type of DBMS based on the relational model introduced by Edgar Codd in 1970. RDBMS stores data in tables (relations) with rows (tuples/records) and columns (attributes/fields); enforces relationships between tables via primary and foreign keys; supports ACID transactions (Atomicity, Consistency, Isolation, Durability); uses SQL as the query language. Examples of RDBMS: MySQL, PostgreSQL, Oracle Database, Microsoft SQL Server, IBM Db2. RDBMS vs DBMS comparison: (1) RDBMS enforces relational integrity via constraints (foreign keys, unique constraints, check constraints); flat DBMS may not. (2) RDBMS supports ACID transactions natively; many NoSQL DBMS relax this. (3) RDBMS uses SQL as the standard query language. (4) RDBMS scales vertically well (better hardware); NoSQL DBMS often scales horizontally better (more servers). For fresher interviews, the key distinction is: DBMS is the broad category, RDBMS is the relational subset that uses tables and SQL.

What is a primary key and what are its properties?

A primary key is a column (or combination of columns) that uniquely identifies each row in a table. Properties of a primary key: (1) Uniqueness - no two rows can have the same primary key value. (2) Not NULL - primary key columns cannot have NULL values (in most databases; SQL Server allows NULL but it's bad practice). (3) Immutability (best practice) - primary key values should not change after insertion, because foreign keys in other tables reference them and changing them would cascade complications. (4) Minimal - the primary key should be minimal; if it is a composite key, no subset of its columns can also uniquely identify rows (otherwise that subset is the candidate key). (5) Stable - primary key values should not be reused or recycled; in production, primary keys are often system-generated (auto-increment integer, UUID, snowflake ID) to guarantee stability. There can be only one primary key per table, but a table can have multiple candidate keys (any column or combination that could be primary key) and the chosen one becomes the primary key. The other candidate keys are alternate keys. The primary key creates a clustered index in SQL Server and InnoDB by default, which physically orders the table data on disk by the primary key.

What is normalization and what are 1NF, 2NF, 3NF, and BCNF?

Normalization is the systematic process of organizing data in a relational database to reduce redundancy and eliminate insertion, update, and deletion anomalies. Each normal form adds a constraint on top of the previous. (1) 1NF (First Normal Form): every column contains atomic (indivisible) values; no repeating groups or arrays within a single row; each row is uniquely identified (typically by a primary key). (2) 2NF (Second Normal Form): 1NF plus every non-key column is fully functionally dependent on the entire primary key - no partial dependencies where a column depends on only part of a composite primary key. (3) 3NF (Third Normal Form): 2NF plus no transitive dependencies where a non-key column depends on another non-key column. (4) BCNF (Boyce-Codd Normal Form): 3NF plus for every functional dependency X -> Y, X must be a superkey - this catches anomalies that 3NF allows in certain edge cases. The practical rule for production databases: design to 3NF or BCNF for transactional data, then denormalize deliberately for reporting or analytics performance. Common fresher question: write the normal forms with a small example showing the violation and the fix.

What are the ACID properties of a transaction?

ACID is an acronym for the four properties that guarantee database transactions are processed reliably: (1) Atomicity - a transaction is all-or-nothing; either all its operations complete successfully and are committed, or none of them do and the database is rolled back to its prior state. Partial completion is impossible. (2) Consistency - a transaction takes the database from one valid state to another valid state; all defined rules (constraints, cascades, triggers) are upheld at commit time. A transaction that violates a constraint is rolled back. (3) Isolation - concurrent transactions do not interfere with each other; each transaction sees a consistent snapshot of the database as if it were the only one running. The level of isolation is configurable (see isolation levels below). (4) Durability - once a transaction commits, its effects survive permanently, even in the face of system crashes, power failures, or other disasters. Durability is typically achieved through write-ahead logging (WAL) - the commit record is flushed to disk before the transaction is acknowledged successful. ACID is the contract between the user and the database management system; violations are bugs. Most RDBMS engines (PostgreSQL, MySQL InnoDB, Oracle, SQL Server) implement ACID by default; many NoSQL databases relax one or more properties for performance.

What are the different transaction isolation levels in SQL?

The SQL standard defines four transaction isolation levels, each providing stronger isolation at the cost of lower concurrency: (1) READ UNCOMMITTED - the lowest level; transactions can see uncommitted changes from other transactions (dirty reads); rarely used in production because it provides almost no isolation. (2) READ COMMITTED - each query sees only data committed before the query began; prevents dirty reads but allows non-repeatable reads (the same query may return different results if re-run in the same transaction because other transactions committed in between). This is the default in PostgreSQL and Oracle. (3) REPEATABLE READ - within a transaction, all reads of the same data return the same value (a snapshot is taken at the start of the transaction); prevents dirty reads and non-repeatable reads but allows phantom reads (new rows can appear when re-querying). This is the default in MySQL/InnoDB. (4) SERIALIZABLE - the strongest level; transactions are executed as if they were serial (one after another); prevents all phenomena including phantom reads; lowest concurrency because the database uses extensive locking or multi-version concurrency control. PostgreSQL uses MVCC (Multi-Version Concurrency Control) for snapshot isolation, which gives better concurrency than locking-based approaches. Common fresher question: explain the three phenomena - dirty read, non-repeatable read, phantom read - and which isolation level prevents each.

What is the difference between a clustered and non-clustered index?

A clustered index determines the physical order of data in a table; the table data is stored on disk sorted by the clustered index key. A table can have only ONE clustered index (because the data can be physically sorted in only one way). In SQL Server and MySQL InnoDB, the primary key constraint creates a clustered index by default; the actual table rows are stored in the B-tree of the clustered index. Reading by clustered index key is very fast (sequential scan within the B-tree); inserting rows in random order can be expensive (may cause page splits). A non-clustered index is a separate structure that contains the index key and a pointer (or row identifier) back to the data row in the clustered index. A table can have many non-clustered indexes (up to 999 in SQL Server, similar limits elsewhere). Non-clustered indexes are faster than full table scans but slower than clustered index lookups because they require an extra lookup to retrieve the full row. Composite indexes (multi-column) work in both clustered and non-clustered forms, but the column order matters - the leftmost column in the index definition should be the most selective (highest cardinality) for query optimization. Practical rule: index columns that appear in WHERE, JOIN ON, and ORDER BY clauses; avoid over-indexing because every index slows down INSERT/UPDATE/DELETE.

What is a deadlock and how does the database handle it?

A deadlock occurs when two or more transactions are each waiting for the other to release a lock, creating a circular wait that prevents any of them from making progress. Example: Transaction A holds lock on Table 1 and is waiting for lock on Table 2; Transaction B holds lock on Table 2 and is waiting for lock on Table 1 - neither can proceed. The database detects deadlocks automatically (most modern RDBMS detect deadlocks within a few seconds via a background deadlock detector thread that examines the lock wait-for graph for cycles). When a deadlock is detected, the database chooses one transaction as the victim (typically the one with the least work done, or based on a configured priority) and rolls it back, releasing its locks so the other transactions can proceed. The rolled-back transaction receives a deadlock error (ORA-00060 in Oracle, error 1213 in MySQL, SQLSTATE 40001 in PostgreSQL) and the application is responsible for retrying it. Prevention strategies: (1) acquire locks in a consistent order across all transactions to prevent circular waits; (2) keep transactions short to minimize lock-hold time; (3) use the lowest sufficient isolation level (e.g., READ COMMITTED instead of SERIALIZABLE); (4) use optimistic concurrency control (version numbers or timestamps) instead of pessimistic locking when possible; (5) add NOWAIT or SKIP LOCKED hints in MySQL/PostgreSQL to fail fast or skip locked rows.

DBMS Interview Questions 2026: Top 40 with Answers

Quick Answer: DBMS Interview Questions 2026

The top 40 DBMS interview questions asked in fresher and experienced interviews at TCS, Infosys, Wipro, Cognizant, and Capgemini in 2026 cover six clusters: RDBMS fundamentals (DBMS vs RDBMS, keys, relationships), normalization (1NF, 2NF, 3NF, BCNF, anomalies, denormalization), SQL (covered in detail in our SQL interview questions guide), transactions and ACID properties (atomicity, consistency, isolation, durability, commit/rollback), concurrency control (locking, deadlocks, isolation levels, phantom reads, dirty reads), and indexing and storage (B-tree, B+ tree, clustered vs non-clustered, query optimization). Most fresher interviews include 5-8 DBMS MCQs plus 1-2 normalization scenarios; experienced interviews add concurrency control, indexing, and database design questions. For the dedicated SQL Q&A set, see the TutorsBot SQL Interview Questions hub.

Why DBMS Is the Second-Highest Weight Topic in Fresher Interviews

DBMS is the second-highest weight topic in service-company fresher interviews (after programming language fundamentals) because: (1) every backend system runs on a database, so you cannot ignore it; (2) DBMS questions test both theoretical understanding (normalization, transactions, ACID) and practical SQL (covered separately); (3) the concepts generalize across all databases (Oracle, MySQL, PostgreSQL, SQL Server) so employers expect new hires to have portable knowledge; (4) DBMS mastery is the foundation for backend, data engineering, and analytics roles. TCS NQT has 5-8 DBMS MCQs; Infosys InfyTQ has 8-10 plus 1-2 normalization scenarios; Cognizant GenC adds concurrency control and indexing questions for the higher tiers.

The 6 Topic Clusters Asked Most Often

Cluster 1: RDBMS Fundamentals

The mandatory baseline. Questions: What is a database (an organized collection of structured information, typically stored electronically and accessed by a computer system)? What is DBMS (the software that manages a database - provides CRUD operations, security, concurrency control, recovery)? What is RDBMS (a DBMS based on the relational model with tables, rows, columns, and ACID transactions; uses SQL)? What are the differences between DBMS and RDBMS (DBMS is the broad category, RDBMS is the relational subset)? What is a key (an attribute or combination of attributes used to identify rows or relationships between tables)? What is a primary key (a column that uniquely identifies each row, NOT NULL, ideally immutable, at most one per table)? What is a foreign key (a column that references the primary key of another table to establish a relationship)? What is a unique key (a column or combination that must be unique but can be NULL; multiple per table)? What is a composite key (a primary key made up of multiple columns)? What are the types of relationships (one-to-one, one-to-many, many-to-many with a junction table)?

Cluster 2: Normalization and Denormalization

Asked at every level. Questions: What is normalization (the systematic process of organizing data to reduce redundancy and eliminate anomalies)? What are 1NF, 2NF, 3NF, and BCNF (each builds on the previous; 1NF = atomic values, 2NF = no partial dependencies, 3NF = no transitive dependencies, BCNF = every determinant is a superkey)? What is a functional dependency (a constraint between attributes - X -> Y means given X, Y is uniquely determined)? What are anomalies (insertion, update, deletion - errors caused by data redundancy)? What is denormalization (deliberately reintroducing redundancy for read performance or operational simplicity - reporting tables, aggregated columns, wide tables)? When should you denormalize (read-heavy reporting, expensive joins, OLAP workloads)? Common live coding: normalize a small schema from 1NF to 3NF, identifying functional dependencies and anomalies. Cognizant GenC and Capgemini love this question.

Cluster 3: SQL (covered separately)

SQL is tested heavily but typically as its own cluster rather than under DBMS. For the full SQL interview prep with 25 curated questions at all difficulty levels, see the TutorsBot SQL Interview Questions hub. The cross-over questions that combine DBMS and SQL: write a CREATE TABLE with all constraint types (PRIMARY KEY, FOREIGN KEY with ON DELETE CASCADE, UNIQUE, CHECK, NOT NULL, DEFAULT), explain the difference between INNER JOIN and OUTER JOIN with execution plans, write a query using window functions with PARTITION BY.

Cluster 4: Transactions and ACID

The classic differentiator. Questions: What is a transaction (a logical unit of work that consists of one or more SQL statements that must be executed atomically)? What are the ACID properties (Atomicity, Consistency, Isolation, Durability - the contract that guarantees reliable transaction processing)? What are the transaction states (active, partially committed, committed, failed, aborted)? What is commit vs rollback (commit makes changes permanent; rollback undoes uncommitted changes)? What is a savepoint (a marker within a transaction to which you can rollback without rolling back the entire transaction)? What is the write-ahead log (WAL - the log file where changes are written before being applied to data files, used to guarantee durability and recovery after crashes)? What is two-phase commit (a protocol for ensuring atomicity across multiple databases in a distributed transaction - phase 1: prepare, phase 2: commit or rollback)? Common scenario: explain what happens at each stage of a transaction when the system crashes in the middle.

Cluster 5: Concurrency Control

Heavy at Infosys Power Programmer, Cognizant GenC Next, and product companies. Questions: What are the three concurrency phenomena - dirty read (reading uncommitted data from another transaction), non-repeatable read (the same query returns different results when re-run in the same transaction), and phantom read (new rows appear when re-querying with the same WHERE clause in the same transaction)? What are the four SQL isolation levels (READ UNCOMMITTED, READ COMMITTED, REPEATABLE READ, SERIALIZABLE) and which phenomena does each prevent? What is the default isolation level in PostgreSQL (READ COMMITTED), MySQL/InnoDB (REPEATABLE READ), Oracle (READ COMMITTED), SQL Server (READ COMMITTED with optional SNAPSHOT)? What is optimistic vs pessimistic locking (optimistic: assume no conflicts, detect at commit via version numbers; pessimistic: lock early, hold to commit)? What is multi-version concurrency control (MVCC) - used by PostgreSQL, MySQL InnoDB, Oracle - each transaction sees a snapshot of the database as it was when the transaction started, allowing readers and writers to not block each other? What are the trade-offs between isolation levels and concurrency (higher isolation = more locks, less concurrency, more contention; lower isolation = faster but more anomalies)?

Cluster 6: Indexing, Storage, and Query Optimization

Asked at Cognizant GenC Next, Infosys Power Programmer, and product companies. Questions: What is an index (a data structure that improves the speed of data retrieval at the cost of additional storage and write overhead)? What is a B-tree (a balanced tree data structure used for most indexes - keeps data sorted, allows logarithmic search, insert, delete)? What is a B+ tree (a variant of B-tree where all data is in leaf nodes and internal nodes only contain keys; better for range queries because leaf nodes are linked)? What is a clustered vs non-clustered index (clustered: defines physical order of data, one per table; non-clustered: separate structure with pointers to data rows, many per table)? What is a hash index (uses a hash function for O(1) equality lookups; no support for range queries)? When should you create an index (on columns that appear in WHERE, JOIN ON, ORDER BY clauses, and have high selectivity - many distinct values)? When should you NOT create an index (small tables, columns that change frequently, columns used only in function expressions)? What is a composite index (an index on multiple columns; the order matters - leftmost prefix rule)? What is the difference between index seek and index scan (seek is highly selective lookup using the index; scan reads the whole table or index range - usually slower)? Common live coding: given a slow query, propose indexes that would speed it up.

Common Fresher Pitfalls to Avoid

Confusing DBMS and RDBMS: DBMS is the broad category (includes hierarchical, network, object, NoSQL); RDBMS is specifically the relational subset using tables and SQL
Confusing primary key and unique key: primary key is one per table, NOT NULL, often the clustered index; unique key is many per table, can have NULL
Forgetting the difference between 2NF and 3NF: 2NF is about partial dependencies on composite keys; 3NF is about transitive dependencies between non-key columns
Confusing READ COMMITTED and REPEATABLE READ: READ COMMITTED allows non-repeatable reads; REPEATABLE READ does not; both prevent dirty reads
Confusing clustered and non-clustered indexes: clustered = physical order, one per table; non-clustered = separate structure, many per table, pointer back to data
Over-indexing: every index slows down INSERT/UPDATE/DELETE because the index must be updated; only index columns that are actually queried
Under-indexing foreign keys: foreign key columns should always be indexed or JOIN ON clauses will be slow
Using SELECT * in production: fetches all columns, breaks if schema changes, wastes bandwidth; explicitly list the columns you need

How to Prepare in 30 Days

The strongest 30-day fresher DBMS interview preparation plan:

Week 1 - Foundations: Work through GeeksforGeeks DBMS section (all 60+ articles), practice 30 normalization questions (start with a small students-marks schema and normalize it step by step from 1NF to 3NF to BCNF)
Week 2 - SQL integration: Combine your DBMS theory with intensive SQL practice (joins, subqueries, window functions) - the TutorsBot SQL Interview Questions hub has 25 curated questions for this
Week 3 - Transactions and concurrency: Practice writing transactions with BEGIN/COMMIT/ROLLBACK, study isolation levels and their phenomena, draw lock-wait graphs for deadlock scenarios
Week 4 - Indexing and design: Use EXPLAIN on slow queries, design schemas for 3-4 real-world scenarios (URL shortener, e-commerce, library, hospital), take 2-3 timed mock tests on HackerRank SQL or LeetCode Database

For a structured, project-based path from DBMS fundamentals to backend developer roles, the TutorsBot Full Stack Development training covers database design, schema management, and API development end-to-end.

Frequently Asked Questions

How many DBMS questions are asked in TCS NQT 2026?

TCS NQT 2026 typically includes 5-8 DBMS questions in the technical section for Digital and Prime profiles. For Ninja profiles, the count is lower (2-4 MCQs). TCS Digital adds harder normalization and concurrency questions for the Java track.

Is DBMS enough to get placed in TCS/Wipro/Infosys?

DBMS alone is rarely enough - companies test a combination of programming language (Java/Python/C++), DBMS, DSA, SQL, and aptitude/reasoning. The strongest fresher profile combines: DBMS theory (normalization, transactions, ACID), SQL proficiency (intermediate to advanced), programming language fluency, DSA (50-100 problems), and 2-3 small projects on GitHub. Cognizant GenC and Capgemini have the heaviest DBMS testing of the major service companies.

What is the salary for DBMS-skilled freshers at these companies in 2026?

Salary bands for freshers with strong DBMS skills in 2026: TCS Ninja ₹3.36 LPA, TCS Digital ₹7-8 LPA, TCS Prime ₹9-11 LPA. Wipro ₹3.5-5 LPA. Infosys (InfyTQ Power Programmer) ₹6.5-9.5 LPA, regular InfyTQ ₹3.6 LPA. Cognizant GenC ₹4.5-6 LPA, GenC Next ₹7-12 LPA. Accenture AASE ₹4.5-6 LPA, AASS ₹6-9 LPA. Capgemini ₹4-6 LPA. HCL ₹3.5-5.5 LPA. Product companies (Flipkart, PhonePe, Razorpay, Cred) pay 2-3x these bands for DBMS-fluent candidates who can also do SQL, programming, and system design.

What is the best resource for DBMS interview preparation?

The strongest resources are: GeeksforGeeks DBMS section (geeksforgeeks.org/dbms - covers everything precisely), StudyTonight DBMS notes (studytonight.com/dbms - clear explanations with examples), TutorialsPoint DBMS (tutorialspoint.com/dbms - quick reference), the TutorsBot SQL Interview Questions hub (25 curated SQL questions at all difficulty levels), and Use The Index Luke (use-the-index-luke.com - SQL performance and indexing deep dive). Combine GeeksforGeeks for breadth, TutorialsPoint for reference, the SQL hub for SQL practice, and Use The Index Luke for indexing and optimization depth.

Can I learn DBMS without learning SQL first?

You can learn DBMS theory (normalization, transactions, ACID, concurrency, indexing) without writing SQL - these are concepts. But to interview well and apply the concepts, you need SQL fluency. The interview pattern is: explain the concept (theory), then write the SQL that demonstrates it (practice). Most fresher interviews expect both - 5-8 DBMS MCQs for theory and 5-8 SQL questions for practice. Learn them in parallel; theory alone or practice alone is incomplete.

Resources and Next Steps

The authoritative sources listed (GeeksforGeeks DBMS, StudyTonight, TutorialsPoint, PostgreSQL documentation, Use The Index Luke) are the canonical references for DBMS fundamentals and interview preparation. For the full SQL interview questions collection with 25 curated questions at all difficulty levels, see the TutorsBot SQL Interview Questions hub. For related interview prep, see our SQL interview questions, Python interview questions, and Normalization in DBMS guides.

#dbms interview questions #dbms interview questions with answers #dbms interview questions for freshers #dbms interview questions for experienced #dbms interview 2026 #rdbms interview questions

← Back to Blog