DBMS for GATE CSE 2027: Important Topics and Patterns

Databases is short on paper and steady in marks. The GATE DBMS important topics for 2027 are SQL queries evaluated on a printed table, conflict serializability, normal forms with decomposition, and B+ tree arithmetic. Almost nothing in DBMS is asked as a pure definition: you compute, then decide. This guide covers what the syllabus names, a priority table, how each topic is asked, the traps and a short plan.

Key takeaways

• The Databases section of the GATE 2027 CS syllabus is two sentences long and identical to 2026.
• Analyst compilations put DBMS at 7–8 marks in most papers since 2019, usually 4–5 questions. These figures are unofficial.
• SQL, serializability and functional dependencies have each appeared in almost every paper since 2009–2010, according to the book's reconstruction of past papers.
• SQL row counts are NAT items. Statement bundles on transactions and decompositions have often been MSQ since 2021.
• Most lost marks come from four traps: NULL logic, > ALL over an empty subquery, recoverability classes and the B+ tree order convention.

What the GATE 2027 DBMS syllabus names

The official GATE 2027 CS syllabus (Section 9) reads:

Databases: ER-model. Relational model: relational algebra, tuple calculus, SQL. Integrity constraints, normal forms. File organization, indexing (e.g., B and B+ trees). Transactions and concurrency control.

The text is unchanged from GATE 2026, and this section has not been reworded since the 2016 revision. Query processing, query optimization and distributed databases are not in the syllabus. By the book's reconstruction, triggers, embedded SQL and RAID have not been asked as stand-alone questions in the 100-mark era either. Confirm the final text at gate2027.iitm.ac.in.

How many marks DBMS carries

The figures below are third-party analyst compilations from gateexam.info, GeeksforGeeks, careers360, goclasses and PW. They are not official, and borderline questions are classified differently by different analysts, so treat each number as ±1–2 marks.

From 2019 the subject has held at 7–8 marks, broken only by 2023 (5). For 2026 Shift 1, GeeksforGeeks reported the DBMS topics as SQL, normalization and transactions; for Shift 2, PW and GeeksforGeeks counted 4 DBMS questions. The 2019 paper, organized by IIT Madras (which also organizes GATE 2027), carried 8–9 DBMS marks and was described by analysts as tricky. For the full picture across subjects, see the GATE CSE subject-wise weightage.

DBMS priority table for GATE 2027

The book ranks every DBMS subtopic with an eight-factor analytical score covering historical and recent frequency, variety of question forms, syllabus wording and fit with NAT or MSQ. The tiers are that ranking, not probabilities of a topic appearing.

On this analysis, a sensible 2027 target is to be ready for one SQL row-count NAT, one normalization item, one transactions item, one B+ tree or indexing NAT and one 1-mark item from keys, ER, relational algebra or constraints. This article gives the tiers. The DBMS chapter of the GATE CSE 2027 book shows the factor scores for all 18 subtopics, a year-by-year concept map and a high-probability question bank with worked solutions.

The most-tested DBMS topics and how they are asked

SQL on a printed instance

Typically three small tables of 4–8 rows and a query with one or two levels of nesting. The answer is a small integer entered as a NAT. The devices setters use, roughly in order of how likely the analysis rates them: a correlated NOT EXISTS for "all" queries (division), a NULL inside a NOT IN subquery, an aggregate subquery compared with a per-row value, > ALL over an empty result, and GROUP BY with HAVING on a count.

The vacuous-ALL device shows how patterns return. The 2012 and 2017 Set 2 papers both used a > ALL comparison whose inner query was empty, five years apart. When the inner result is empty, > ALL is true for every outer row and > ANY is false.

Conflict serializability and schedule classes

Two shapes dominate. Either "which serial schedule is S equivalent to", with three transactions and 6–8 operations, or an MSQ listing four properties of one schedule: conflict serializable, recoverable, cascadeless, strict. A counting variant, the number of conflict-serializable interleavings of two transactions, appeared in 2017 Set 1. The analyst reconstruction finds a serializability or recoverability item in every paper since 2019.

Normal forms and decomposition

A schema with 4–6 attributes and 3–4 functional dependencies. Find every candidate key first, then decide the highest normal form, or test a decomposition MSQ: is it lossless, is it dependency preserving, is each component in BCNF. The "3NF but not BCNF" characterisation was asked in 2018 and 2019.

B+ tree arithmetic

Block size, key size and pointer sizes give the maximum keys in an internal node or a leaf, followed by levels or block accesses for N records. The same template was reused in 2010, 2015 and 2020.

A short example: block 1,024 bytes, key 12 bytes, block pointer 8 bytes. An internal node with p pointers needs p × 8 + (p − 1) × 12 ≤ 1,024, so 20p ≤ 1,036 and p = 51. A leaf stores record pointers plus one sibling pointer, so it uses a different inequality. Mixing the two is a common error.

Tier 2 regulars

Candidate keys are counted with attribute closure. Relational algebra questions often use a self-join with rename and a difference to express "not the maximum" or "more than every". ER-to-table mapping (minimum number of relations) has had a long gap since its last strong appearance in 2018, which is why the analysis still rates it High. NULL questions test COUNT(*) against COUNT(column) and grouping on a nullable column.

Common DBMS traps

• NOT IN with a NULL. If the subquery returns a NULL, NOT IN returns no rows. NOT EXISTS gives the intended meaning.
• Empty subqueries. > ALL, >= ALL and < ALL are true; > ANY and IN are false; NOT EXISTS is true.
• Aggregates and NULL. COUNT(*) counts rows. COUNT(column) and the other aggregates skip NULLs. GROUP BY puts all NULLs in one group.
• Conflicts. Read–read is not a conflict, and operations of the same transaction never conflict.
• Recoverability. A schedule where the reader commits after the writer it read from is recoverable. Cascadeless forbids reading uncommitted data; strict forbids reading or writing it.
• 2PL guarantees. Basic 2PL gives conflict serializability but prevents neither deadlock nor cascading aborts.
• F versus F+. Test losslessness and dependency preservation using F+, not only the listed F.
• All-prime schemas. If every attribute is prime, the schema is already in 3NF. Only BCNF needs checking.
• B+ tree order. Check whether order means children or keys, include the leaf's sibling pointer, and add one access for the data block.
• Keys. Attributes that appear on no right-hand side belong to every candidate key.

These traps and their counterparts in other subjects are collected in common mistakes that cost marks in GATE CSE.

A short DBMS preparation plan

1. Master three procedures first. Attribute closure, the precedence graph and row-by-row SQL evaluation. According to the book's recurrence analysis, a candidate who runs these three flawlessly under time pressure has historically secured 4–6 of the DBMS marks.
2. Practise SQL on tiny tables with NULLs. Write the intermediate result after each clause, in the order FROM, WHERE, GROUP BY, HAVING, SELECT. Never answer by guessing what the query intends.
3. Work the FD chain in one sitting. Keys, then highest normal form, then a decomposition check. Repeat on schemas with cyclic dependencies.
4. Classify schedules against all four properties. For every practice schedule, decide conflict serializability, recoverability, cascadelessness and strictness together, the way an MSQ asks. The MCQ, MSQ and NAT strategy guide covers how to mark such items.
5. Drill B+ tree capacity and height. Then do one insertion sequence by hand, noting that a leaf split copies a key up while an internal split pushes it up.
6. Finish Tier 3 and Tier 4 in one pass. Index types, canonical cover, timestamp ordering and log-based recovery are mostly 1-mark statements.

For revision, the GATE CSE formula sheet has the core DBMS formulas, and the most repeated topics in GATE CSE shows where DBMS sits among the highest-priority concepts across the paper.