Computer Science

DBMS Master Revision Handbook

12 September 2026

Share this study resource

Help your exam peers & study groups prepare better

WhatsApp Telegram 𝕏

🗄️ DBMS / SQL — Master Revision Handbook

🚨 15+ Answer Key Bugs Detected & Fixed 📊 467 MCQs Analyzed 🎯 Exam-Ready | Hindi ⭐ High-Yield Only

🚨 EXAM TRAPS & ANSWER KEY BUGS — सर्वोच्च प्राथमिकता

🚨 Answer Key Bugs — गलत उत्तर चिह्नित हैं (Book WRONG → Correct सही)

विषय	प्रश्न (संक्षिप्त)	Book (❌)	Correct (✅)	कारण
Basic	Data कहाँ store होता है?	(b) DBMS में	(d) Database में	DBMS = Manager, Database = Storage. दोनों अलग हैं।
DB Models	सबसे पुराना DB model?	(c) Network	(a) Hierarchical	Hierarchical पहले आया (IBM IMS 1960s), फिर Network।
RDBMS	'Attribute' क्या होता है?	(b) Tuple (Row)	(a) Column	Attribute = Column/Field। Tuple = Row। सबसे ज़्यादा confuse होने वाला trap।
RDBMS	Relational Model किससे संबंधित?	(b) Structure+Integrity	(c) Both a and b (सभी)	Data Manipulation भी Relational Model का हिस्सा है।
Keys	Student table में Primary Key?	(b) Dept	(c) ID	Dept एक से ज़्यादा students में share होता है — PK नहीं बन सकता।
Keys	Rows क्या कहलाती हैं?	(a) Relation	(b) Tuples	Relation = Table (पूरी), Tuple = Row। Book बिल्कुल गलत।
Keys	FK — Referencing vs Referenced?	Referencing	Referenced	FK रखने वाली table = Referencing; FK जहाँ point हो = Referenced।
Rela. Types	Relational DB में किस type के relations?	(c) Flat	(b) Normalized	Flat-file और Relational DB अलग concepts हैं।
Rel. Degree	Relationship degree के types?	(a) Binary only	(d) All of above	Unary, Binary, Ternary, N-ary — सभी valid degree types हैं।
DB Design	E-R Model किसने दिया?	Dr. Edgar F. Codd	Peter Chen (1976)	🚨 MAJOR TRAP: Codd = Relational Model, Peter Chen = ER Model।
Design	Relational Model में tables को क्या कहते?	(b) Records	(c) Relations	"Relational" शब्द ही answer देता है — tables = Relations।
DDL/DML	DDL नहीं है कौनसा? (SQL में)	(a) RENAME only	(d) UPDATE	RENAME = DDL (structure change)। UPDATE = DML (data change)।
DDL/DML	DML नहीं है कौनसा?	(b) CREATE	(d) TRUNCATE	TRUNCATE = DDL। CREATE भी DDL। Book ने गलत option mark किया।
NoSQL	MongoDB किसमें लिखा गया?	(a) JavaScript	(b) C++	MongoDB का core engine C++ में है। Query language JS है।
NoSQL	DELETE rows के लिए SQL command?	(d) REMOVE	(a) DELETE	REMOVE कोई valid SQL command नहीं है।
NoSQL	NoSQL databases को refer किया जाता है?	(b) Non-relational	(c) "Not Only SQL"	NoSQL = "Not Only SQL" — official correct expansion।
Manipulating	Record management primitives?	(c) Look-up only	(d) All of above	Add, Delete, Look-up — तीनों primitive operations हैं।
GRANT/REVOKE	GRANT/REVOKE किसमें आते हैं?	कई books में DDL	DCL (Data Control Language)	🚨 Mock Test bug: GRANT/REVOKE = हमेशा DCL, कभी DDL नहीं।

⚠️ सबसे ज़्यादा Confuse होने वाले TRAPS

Attribute vs Tuple — सबसे ज़्यादा गलती
Attribute = Column/Field (vertical)
Tuple = Row/Record (horizontal)
Relation = Table (complete 2D)
🧠 A-ttribute = A column। T-uple = T-able row।

Referenced vs Referencing Relation
Referencing relation = जिसमें FK रखा जाता है (doing the referencing)
Referenced relation = जहाँ PK होती है जिसे FK point करता है
🧠 "Who has the FK" = Referencing। "Who has the PK" = Referenced।

ACID — Atomicity vs Durability Timeline Trap
Transaction के दौरान failure → Atomicity (Rollback)
Transaction Commit के बाद crash → Durability (Persist)
🧠 "Transaction सफल के बाद भी safe" = Durability (not Atomicity!)

ER Model — Aggregation Trap
"Relationships are treated as higher-level entities" पढ़ते ही Aggregation चुनो
Generalization ≠ Aggregation
🧠 Entity का एकीकरण = Generalization; Relationship का एकीकरण = Aggregation

WHERE vs HAVING — Aggregate Function
WHERE AVG(Salary) > 5000 ❌ Error देगा
HAVING AVG(Salary) > 5000 ✅ सही
🧠 W comes before H। WHERE = Worker level, HAVING = Head/Group level

UNION vs UNION ALL — Performance
UNION ALL > UNION (performance में तेज़)
क्योंकि UNION ALL sorting नहीं करता
🧠 UNION ALL = All Includes (बिना छानबीन)। UNION = Unique (extra work)

Codd vs Peter Chen — कभी मत भूलो
Edgar F. Codd → Relational Model (1970)
Peter Chen → E-R Model (1976)
Charles Bachman → First DBMS (IDS)
🧠 C for Codd = C for Columns/Relational। Chen = E-R = Entity Relationship

TRUNCATE — DDL या DML?
TRUNCATE = DDL (हमेशा)
Auto-committed है, Rollback नहीं होता
DELETE = DML, Rollback हो सकता है
🧠 TRUNCATE = Table Reset (structure change style)।

MySQL Full Outer Join Trap — "FULL OUTER JOIN हर Database में support होता है" → गलत
MySQL में FULL OUTER JOIN directly support नहीं होता। Method: LEFT JOIN ... UNION ... RIGHT JOIN

📚 CORE CONCEPTS — Rapid Revision

1. DBMS — मूल परिचय

Term	Full Form / अर्थ
DBMS	Database Management System
RDBMS	Relational DBMS (tables में)
DDL	Data Definition Language
DML	Data Manipulation Language
DCL	Data Control Language
TCL	Transaction Control Language
DQL	Data Query Language (SELECT)

Important Persons	Contribution
Charles Bachman	World's First DBMS (IDS, 1960)
Edgar F. Codd	Relational Model (1970), 12 Rules
Peter Chen	E-R Model (1976)
Carlo Strozzi	"NoSQL" term (1998)

🧠 B→C→P: Bachman (First DBMS) → Codd (Relational) → Peter Chen (ER Model)

2. DB Models — तुलना ⭐⭐

Model	Structure	Relations	Example	कब efficient
Hierarchical	Tree (oldest)	One-to-Many, ONE parent per child	IBM IMS	Small transactions + Large data
Network	Graph	Many-to-Many, MULTIPLE parents allowed	CODASYL	Complex relationships
Relational	Tables (Relations)	All types via FK/joins	MySQL, Oracle	General purpose (most used)
Object-Oriented	Objects	Class hierarchy	ObjectDB	Complex data types

🧠 Hierarchical = Family Tree (एक parent per child)। Network = Social Network (multiple parents)। "H before N" in alphabet = Hierarchical before Network (history wise)

3. ANSI/SPARC Three-Schema Architecture ⭐⭐

External Schema
= User View (partial)
External users जो देखते हैं
Multiple views हो सकती हैं

Conceptual Schema
= Complete Logical View
DBA के लिए entire DB का view
एक ही होती है

Internal Schema
= Physical Storage Structure
Disk पर कैसे store होगा
एक ही होती है

⚠️ Trap: 'Implementation Schema' — यह EXIST नहीं करता। Internal Schema = ANSI/SPARC में है, Implementation नहीं।
🧠 ECI = External, Conceptual, Internal — तीन real schemas।

4. Keys — Complete Summary ⭐⭐⭐

Key Type	परिभाषा	Rules
Primary Key (PK)	Unique + NOT NULL — एक table में एक ही	NULL❌, Duplicate❌
Foreign Key (FK)	दूसरी table की PK को refer करे	Referential Integrity
Candidate Key	Uniquely identify कर सके — सभी possible keys	एक table में multiple हो सकती हैं
Alternate Key	Candidate Keys जो PK नहीं बनीं	—
Unique Key	Unique values ensure करे	NULL✅ (one allowed), Duplicate❌
Composite Key	दो या अधिक columns मिलकर PK	—
Super Key	Candidate Key के superset (extra attributes के साथ)	—

Primary = ONE per table, NO NULL | Unique = ONE NULL allowed | Candidate Keys = Multiple | All Candidate keys can uniquely identify rows

5. DDL / DML / DCL / TCL ⭐⭐⭐

Category	Full Form	Commands	Auto-Commit?	Rollback?
DDL	Data Definition Language	CREATE, ALTER, DROP, TRUNCATE, RENAME	✅ हाँ	❌ नहीं
DML	Data Manipulation Language	SELECT, INSERT, UPDATE, DELETE	❌ नहीं	✅ हाँ
DCL	Data Control Language	GRANT, REVOKE	✅ हाँ	❌ नहीं
TCL	Transaction Control Language	COMMIT, ROLLBACK, SAVEPOINT	—	—

🚨 Bug: GRANT/REVOKE = हमेशा DCL। कई mock tests में इसे DDL में डाल देते हैं — गलत।

TRUNCATE vs DELETE vs DROP ⭐⭐

Feature	TRUNCATE	DELETE	DROP
Category	DDL	DML	DDL
क्या हटाता है	सभी Rows (data)	Specific rows (WHERE से)	पूरी Table structure + data
Rollback	❌ नहीं	✅ हाँ	❌ नहीं
Log	Minimal logging	Full row-level logging	—
Speed	तेज़	धीमा (per row)	—
Table Structure	बना रहता है	बना रहता है	हट जाता है

💻 SQL — Complete Exam Reference

SQL Execution Order ⭐⭐⭐ — सबसे ज़्यादा पूछा जाता है

FROM → JOIN → WHERE → GROUP BY → HAVING → SELECT → DISTINCT → ORDER BY → LIMIT/TOP

FROM = Source table पहचानना
WHERE = Individual rows filter (Aggregates नहीं)
GROUP BY = Groups बनाना
HAVING = Groups filter (Aggregates यहाँ)

SELECT = Columns चुनना (Alias यहाँ define होता है)
ORDER BY = Sort करना (SELECT alias यहाँ use हो सकता है)
Alias को WHERE में USE नहीं कर सकते — SELECT बाद में execute होता है

WHERE vs HAVING ⭐⭐

Feature	WHERE	HAVING
Filter करता है	Individual Rows (Row-level)	Groups (Group-level)
कब execute होता है	GROUP BY से पहले	GROUP BY के बाद
Aggregate Functions	❌ नहीं (Error)	✅ हाँ (SUM, COUNT, AVG...)
बिना GROUP BY	अकेले काम करता है	पूरी table को 1 group मानता है

SQL Execution Order Trick: Fw Gh So
FROM → WHERE → GROUP BY → HAVING → SELECT → ORDER BY

JOINs ⭐⭐⭐

JOIN Type	Returns	Unmatched Rows
INNER JOIN	केवल Common (matching) records	पूरी तरह गायब
LEFT OUTER JOIN	Left table सभी + Right matching	Right side → NULL
RIGHT OUTER JOIN	Right table सभी + Left matching	Left side → NULL
FULL OUTER JOIN	दोनों tables सभी records	दोनों तरफ NULL possible
CROSS JOIN	Cartesian Product (m × n)	कोई condition नहीं
SELF JOIN	Table खुद से join	Alias अनिवार्य (e1, e2)
NATURAL JOIN	Common column name पर auto join	No common column → CROSS JOIN

🧠 Keyword trap: "even if doesn't exist / all records regardless" → LEFT या RIGHT JOIN चुनो (जो 'All' वाली table उस side हो)

⚠️ INNER JOIN Maximum rows ≤ m × n (Cartesian Product)। कभी exceed नहीं होगा।
MySQL में FULL OUTER JOIN नहीं होता → LEFT JOIN + UNION + RIGHT JOIN

Set Operations ⭐⭐

Operator	Output	Duplicates	Performance	Commutative?
UNION	दोनों की सभी unique rows	Remove करता है	Slow (sorting)	✅ A∪B = B∪A
UNION ALL	दोनों की सभी rows	रखता है (Keeps)	Fast ⚡	✅ A∪B = B∪A
INTERSECT	केवल Common rows	Remove करता है	Slow	✅ A∩B = B∩A
MINUS/EXCEPT	पहले में से दूसरे को घटाओ	Remove करता है	Slow	❌ A-B ≠ B-A

Set Operations की अनिवार्य शर्तें: ① दोनों queries में columns की संख्या समान ② Corresponding columns का data type compatible ③ Column names समान होना ज़रूरी नहीं (output = पहली query के column names)
ORDER BY केवल पूरी query के अंत में एक बार (individual SELECTs के साथ नहीं)

Subquery / Nested Query ⭐⭐

Inner Query पहले execute होती है
Inner Query result = Outer Query को देती है
WHERE या HAVING में use होती है
ANY = कम से कम एक से match (OR)
ALL = सभी से match (AND)
EXISTS = TRUE/FALSE (row exist करे तो TRUE)

🧠 ANY vs ALL:
ANY = At least one (OR condition)
ALL = Every single one (AND condition)
WHERE salary > ALL(subquery) = सबसे ज़्यादा MAX से भी ज़्यादा
WHERE salary > ANY(subquery) = कम से कम MIN से ज़्यादा

NOT IN vs NOT EXISTS — Null Trap ⭐

⚠️ Important: अगर Subquery में NULL आए तो:
NOT IN → पूरा result empty/NULL हो जाता है (NULL से compare = unknown)
NOT EXISTS → NULL का कोई ऐसा effect नहीं पड़ता (safer)

Indexing ⭐⭐

Feature	With Index	Without Index
Read (SELECT)	✅ Fast — O(log N)	❌ Slow — Full Table Scan O(N)
Write (INSERT/UPDATE)	❌ Slower (index update)	✅ Faster
Storage	❌ Extra space लेता है	✅ कम space

🧠 Index = Read Fast, Write Slow | Index = Extra Storage (reduces nothing)

File Organization: Hash vs B+ Tree ⭐⭐

Feature	Hash Organization	B+ Tree Organization
Structure	Hash Table / Buckets	Balanced Search Tree (Leaf nodes linked)
Best for	Exact Match (WHERE id = 10)	Range Query (BETWEEN, >, <, ORDER BY)
Sorting/ORDER BY	❌ नहीं (unsupported)	✅ हाँ (efficient)
Time Complexity	O(1) average	O(log N)

⚠️ Trap: "Hash O(1) है तो हमेशा बेहतर" — गलत। Range queries और ORDER BY में B+ Tree ही काम करता है।
🧠 Hash = Handshake (Exact person) | B+ Tree = Between/Bigger-Smaller (Range)

🔄 ACID Properties & Transactions

ACID Properties ⭐⭐⭐

Property	Core Meaning	Exam Keyword	Responsible Component
Atomicity	All or Nothing — पूरा या कुछ नहीं	Rollback / Transaction के दौरान failure	Recovery Manager
Consistency	DB के नियम (constraints) कभी नहीं टूटेंगे	Valid State / Integrity Constraints	User / Programmer
Isolation	Concurrent transactions एक-दूसरे को disturb नहीं करते	Dirty Read / Concurrent execution	Concurrency Control (Locks)
Durability	Commit के बाद crash में भी data safe	System Failure / Persist / Disk	Recovery Manager (Log)

🚨 Timeline Trap:
Transaction के दौरान failure → Rollback → Atomicity
Transaction Commit के बाद crash → Data persist रहे → Durability
"Dirty Read problem" = Isolation का उल्लंघन

Concurrency Control Anomalies ⭐

Anomaly	क्या होता है	Solution
Dirty Read	Uncommitted data को दूसरा पढ़ ले	Isolation / Locking
Non-Repeatable Read	Same query, अलग result (update के बाद)	Higher Isolation Level
Phantom Read	Same query, नई rows appear हों (insert के बाद)	Serializable Isolation
Lost Update	दो transactions एक-दूसरे का update overwrite करें	Locking

Cascading Rollback ⭐

Cascading Rollback: एक Transaction (T₁) fail → उसका Dirty Read करने वाले सभी Transactions (T₂, T₃...) भी Rollback होंगे → Chain Reaction
कारण: Dirty Read (Uncommitted Dependency)
बचाव: Cascadeless Schedule — कोई भी transaction, दूसरे का uncommitted data नहीं पढ़ सकता
Hierarchy: Strict ⊂ Cascadeless ⊂ Recoverable
🧠 Cascading = Waterfall (ऊपर गिरा तो नीचे सब गिरे)

Transaction States

Active → Partially Committed → Committed (success) | Active → Failed → Aborted (rollback)

📐 ER Model & Database Design

ER Diagram Symbols ⭐⭐⭐

Symbol	Represents	Example
Rectangle (आयत)	Entity	Student, Employee
Double Rectangle (दोहरी आयत)	Weak Entity	Dependent (of Employee)
Diamond (हीरा)	Relationship	Works-In
Ellipse (दीर्घवृत्त)	Attribute	Name, Age
Double Ellipse	Multivalued Attribute	Phone Numbers
Dotted Ellipse	Derived Attribute	Age (from DOB)
Underlined in Ellipse	Key Attribute	Student_ID
Dashed Rectangle	Aggregation	Relationship acting as Entity

🧠 Diamond = Relationship (diamond ring connects two people) | Rectangle = Entity | Ellipse = Attribute | Double = Weak/Multivalued

Generalization vs Specialization vs Aggregation ⭐⭐

Concept	Approach	क्या होता है	ER Symbol
Generalization	Bottom-Up ↑	कई entities → एक Higher Entity	Triangle (ISA)
Specialization	Top-Down ↓	एक entity → कई sub-entities	Triangle (ISA)
Aggregation	Abstraction	Relationship → Higher Entity बन जाए	Dashed Rectangle

⚠️ "Relationships treated as higher-level entities" → Aggregation (Generalization नहीं!)
🧠 Ground→Sky = Generalization (Bottom-Up) | Sky→Ground = Specialization (Top-Down) | RA = Relationship to Aggregation

Strong Entity vs Weak Entity

Strong Entity:
✅ अपनी Primary Key होती है
✅ स्वतंत्र रूप से exist करती है
Symbol: Single Rectangle

Weak Entity:
❌ Primary Key नहीं होती
❌ Strong entity पर depend करती है
Symbol: Double Rectangle (दोहरी)
Discriminator/Partial key होती है

Normalization ⭐

Normal Form	Condition
1NF	कोई Repeating Groups नहीं, Atomic values (Multivalued attribute नहीं)
2NF	1NF + No Partial Dependency (Non-key attributes पूरी PK पर depend)
3NF	2NF + No Transitive Dependency
BCNF	3NF + हर determinant एक Candidate Key होनी चाहिए

🌐 NoSQL Technology

NoSQL — परिचय ⭐⭐

NoSQL = "Not Only SQL" (NOT "Non-relational" ❌)
Schema-free — पहले structure define नहीं करना
Join-free — tables join नहीं होती
Big Data के लिए बना — Unstructured data
High Availability through distributed architecture
SQL से ज़्यादा flexible और faster development

NoSQL Types	Examples
Document	MongoDB, CouchDB
Key-Value	Redis, DynamoDB (simplest)
Column	Cassandra, HBase
Graph	Neo4j

🧠 D-K-C-G: Document, Key-value, Column, Graph

NoSQL vs SQL (RDBMS) ⭐

Feature	SQL (RDBMS)	NoSQL
Schema	Fixed (predefined)	Flexible / Schema-free
ACID	✅ Full ACID compliance	❌ Generally BASE (eventually consistent)
Scaling	Vertical (scale up)	Horizontal (scale out)
Data Model	Tables (Relations)	Document/Key-value/Graph/Column
Joins	✅ Complex joins	❌ Usually no joins
Storage	Row-based	Document=JSON/BSON; KV=Key-Value pairs

MongoDB Specifics ⭐

MongoDB Term	RDBMS Equivalent
Collection	Table
Document	Row / Record
Field	Column / Attribute
_id	Primary Key

MongoDB core engine = C++ में लिखा है (JavaScript नहीं ❌ — JS shell के लिए है)

NoSQL Answer Key Bugs ⭐

🚨 Confirmed Bugs in NoSQL Questions:
① Impala = Cloudera का NoSQL tool (HBase नहीं) ✅
② NoSQL का aim = Non-relational data manage करना (broader claims गलत)
③ NoSQL से data extract = Query लिखनी और run दोनों करनी होती है
④ NoSQL = "Not Only SQL" (Non-relational नहीं — यह limited definition है)
⑤ TRUNCATE = DDL (book कई बार DML में count करती है — गलत)

🏗️ Architecture & Advanced Concepts

Database Architecture Types ⭐

Architecture	Description	Use Case
1-Tier	Standalone — DB, application, user एक ही machine	Local development
2-Tier	Client-Server — Client application directly DB server से	Traditional enterprise apps
3-Tier	Client + Application Server + DB Server	Web applications

Server Processes ⭐

Process	काम
Server Process	Queries execute करना, results client को return करना
Lock Manager	Lock grant, release, Deadlock detection
DB Writer Process	Buffer से disk पर data write करना
Log Writer Process	Transaction log write करना

Transaction-server system = Query-server system (Server side SQL process करता है)

Parallel Databases ⭐

Coarse-Granularity Parallelism:
Few processors (general-purpose)
Large data chunks

Fine-Granularity Parallelism:
Many specialized processors
Tiny data chunks
Degree of Parallelism = एक operation पर कितने parallel servers

Database Sharding vs Federation ⭐

Feature	Sharding	Federation
Nature	Homogeneous (same schema)	Heterogeneous (different schemas)
What splits	एक DB की Rows — अलग servers पर	कई independent DBs को एक window से जोड़ना
Autonomy	❌ Shards independent नहीं	✅ Each DB fully autonomous
Coupling	Tightly coupled	Loosely coupled

🧠 Federation = Freedom (हर DB अपना boss) | Sharding = Shredding (एक DB को काट दिया rows में)

Important DBMS Facts — Quick Recall ⭐⭐

Topic	Key Fact
Data Dictionary	Metadata के बारे में metadata = System Catalog। DB के सभी objects describe करता है।
Buffer	Disk-to-Memory transfers minimize करता है। RAM में data cache।
DBA	Database Performance के लिए सबसे ज़्यादा जिम्मेदार
Data Mining	Database पर directly काम करके patterns/knowledge निकालना
Dynaset	MS Access में query का result = Dynamic Set of records
Google	Search Engine है, DBMS नहीं (DBMS = Oracle, MySQL, DB2, MS Access)
Cloud: On-Demand	Human interaction के बिना resources provision करना = On-Demand Self-Service
Data Hierarchy	Bit → Byte → Field → Record → File → Database
Master File	Original/Primary permanent data file of a system
Embedded SQL	SQL code inside 3GL programs (C, Java, COBOL)

📌 One-Line Revision — Last Minute Rapid Recall

Topic	One-Line Fact
DBMS vs Database	DBMS = Manager/Software, Database = Storage. दोनों अलग।
Hierarchical (Oldest)	IMS (IBM, 1960s)। Tree। One parent per child।
Network	Graph। Multiple parents per child। CODASYL।
Relational	Tables = Relations। Set concept। SQL। Codd ने बनाया।
Primary Key	Unique + NOT NULL। एक table में एक ही।
Foreign Key	दूसरी table की PK। Parent-Child relationship।
Unique Key	Unique। एक NULL allowed।
DDL	CREATE ALTER DROP TRUNCATE RENAME — Structure बदलता है।
DML	SELECT INSERT UPDATE DELETE — Data बदलता है।
DCL	GRANT REVOKE — Permissions control।
TCL	COMMIT ROLLBACK SAVEPOINT — Transaction control।
TRUNCATE	DDL। Auto-committed। Rollback नहीं।
WHERE	Rows filter। GROUP BY से पहले। Aggregates नहीं।
HAVING	Groups filter। GROUP BY के बाद। Aggregates ✅।
INNER JOIN	केवल matching records।
LEFT JOIN	Left table सभी + Right NULL।
CROSS JOIN	Cartesian Product = m × n।
UNION	Duplicates हटाकर combine। Slow।
UNION ALL	Duplicates रखकर combine। Fast।
INTERSECT	केवल common rows।
MINUS/EXCEPT	पहले में से दूसरे को घटाओ। Non-commutative।
Subquery	Inner query पहले run। Outer query को result देती है।
Index	Read fast, Write slow। Extra storage।
B+ Tree	Range queries, ORDER BY। O(log N)।
Hash	Exact match। O(1)। Range नहीं।
Atomicity	Transaction दौरान failure → Rollback। All or Nothing।
Durability	Commit के बाद crash → Data safe। Disk पर permanent।
Isolation	Concurrent transactions interfere नहीं करते।
Dirty Read	Uncommitted data पढ़ना → Isolation violation।
Cascading Rollback	Dirty Read → एक fail → सब fail। Cascadeless से बचाव।
Generalization	Bottom-Up। कई entities → एक। ISA triangle।
Specialization	Top-Down। एक entity → कई। ISA triangle।
Aggregation	Relationship → Entity। Dashed rectangle।
Weak Entity	No Primary Key। Depends on Strong entity। Double rectangle।
Schema	DB की logical structure। Instance = actual data at a time।
NoSQL	"Not Only SQL"। Schema-free। Big Data। 4 types।
MongoDB	Document DB। C++ में written। Collection=Table, Document=Row।
Sharding	एक DB को rows में split। Homogeneous। Autonomy खोना।
Federation	कई DBs → एक window। Heterogeneous। Autonomy बनाए रखना।
SQL Execution Order	FROM→WHERE→GROUP BY→HAVING→SELECT→ORDER BY
Alias in WHERE	SELECT alias को WHERE में use नहीं कर सकते (execution order)।

🧠 All Memory Tricks — एक जगह

Codd vs Chen: C for Codd = C for Columns/Relational (1970). Chen = ER = Entity Relationship (1976). B for Bachman = B for Beginning (1st DBMS).

Execution Order (Fw Gh So): FROM → WHERE → GROUP BY → HAVING → SELECT → ORDER BY

ACID: A=Atomic(All/Nothing), C=Consistency(Valid), I=Isolated(No disturbance), D=Disk(Durable after commit)

Keys: Primary = ONE per table, Perfect (No NULL). Unique = U = 1 NULL allowed. Foreign = Foreign table की PK।

H before N: Hierarchical (oldest, 1960s) → Network → Relational. "H comes before N in alphabet"

DDL = Design, DML = Data, DCL = Cops (Security), TCL = Time-management

JOIN Selection: "all records regardless / even if doesn't exist" → LEFT JOIN। m×n = CROSS JOIN।

Diamond ♦ = Relationship (diamond ring connects two). Rectangle = Entity. Ellipse = Attribute.

ANY = at least one (OR), ALL = every single one (AND). "ALL" से बड़ा = MAX से भी बड़ा।

Index = Read Fast, Write Slow (किताब के पीछे index से पन्ना ढूंढना fast; बीच में नया पन्ना जोड़ना slow)

NoSQL Types D-K-C-G: Document, Key-value, Column, Graph

Cascading = Waterfall: ऊपर गिरा तो नीचे सब गिरे। बचाव = Dirty Read मत करो।

Federation = Freedom (हर DB अपना boss) | Sharding = Shredding (rows में काट दो)

Attribute = Column, Tuple = Row: A-ttribute = A column। T-uple = T-able row।

WHERE=Worker(Row-level), HAVING=Head(Group-level). W comes before H alphabetically।

NOT IN vs NOT EXISTS: NOT IN + NULL = empty result। NOT EXISTS = safer with NULLs।

🗄️ DBMS Master Revision Handbook | 467 MCQs Analyzed | 18 Answer Key Bugs Fixed | Rajasthan Exam Ready

Rajasthan Exam Twister

Comprehensive Preparation for Rajasthan Exams

Review detailed blog breakdowns, previous year papers, and topical revision handbooks.

Browse Blogs