SQL Server 2025 Cheat Sheet

Updated 2026-04-21

Next Topic: SQL Server Management Studio (SSMS) Cheat Sheet

🧠Study flashcards on this topic180 cards · spaced repetition→

SQL Server 2025 is Microsoft's flagship relational database management system (RDBMS), a comprehensive platform for enterprise data storage, processing, and analytics that runs on both Windows and Linux. This release, generally available since November 2025, introduces transformative capabilities including native AI integration with vector search and DiskANN indexing, a native JSON data type with binary storage and indexing, regular expression functions, and significant performance improvements through intelligent query processing. Understanding SQL Server's rich feature set—from core T-SQL syntax and indexing strategies to advanced capabilities like temporal tables, Always On availability, and in-memory OLTP—is essential for building scalable, high-performance database solutions.

What This Cheat Sheet Covers

This topic spans 33 focused tables and 290 indexed concepts, 180 flashcards. Below is a complete table-by-table outline of this topic, spanning foundational concepts through advanced details.

Table 1: Core Data TypesTable 2: Query FundamentalsTable 3: JOIN OperationsTable 4: Aggregate FunctionsTable 5: Window FunctionsTable 6: String FunctionsTable 7: Regular Expression Functions (New in 2025)Table 8: Fuzzy String Matching (Preview in 2025)Table 9: Date and Time FunctionsTable 10: Conditional LogicTable 11: SubqueriesTable 12: Common Table Expressions (CTEs)Table 13: Data ModificationTable 14: IndexesTable 15: ConstraintsTable 16: TransactionsTable 17: Isolation LevelsTable 18: Stored Procedures and FunctionsTable 19: Error HandlingTable 20: JSON Functions (Enhanced in 2025)Table 21: XML FunctionsTable 22: Data Transformation OperatorsTable 23: Table PartitioningTable 24: Temporal Tables (System-Versioned)Table 25: High Availability and Disaster RecoveryTable 26: Performance and Query OptimizationTable 27: Backup and RecoveryTable 28: Security and PermissionsTable 29: In-Memory OLTPTable 30: Advanced T-SQL FeaturesTable 31: Vector Search and AI Integration (New in 2025)Table 32: External Integration and Database Engine (New in 2025)Table 33: System Functions and Metadata

Quick IndexSubscribe to unlock

A jump-to index of every table row in this cheat sheet.

Mind MapSubscribe to unlock

An interactive map of every table and concept in this topic.

Table 1: Core Data Types

Choosing the right data type is the first decision in any schema, and it shapes storage size, precision, and performance for the life of the table. Alongside the familiar integers, strings, and dates you'll find the two headline additions in 2025 — a native binary JSON type and a VECTOR type for AI embeddings — so it's worth knowing which workhorse to reach for and where the new options fit.

Type	Example	Description
INT	`product_id INT`	• Standard 4-byte integer ranging from -2,147,483,648 to 2,147,483,647 • the most common whole number type.
BIGINT	`order_total BIGINT`	8-byte integer supporting extremely large values (-9,223,372,036,854,775,808 to 9,223,372,036,854,775,807).
VARCHAR(n)	`name VARCHAR(100)`	• Variable-length string storing up to `n` characters (max 8,000) • uses only space needed plus 2 bytes overhead.
NVARCHAR(n)	`title NVARCHAR(200)`	• Unicode variable-length string supporting international characters • 2 bytes per character, max 4,000 characters.
DECIMAL(p,s)	`price DECIMAL(10,2)`	• Fixed-precision numeric where `p` is total digits (1-38) and `s` is decimal places • essential for financial data.
DATETIME2	`created_at DATETIME2`	• High-precision datetime with 100-nanosecond accuracy • range 0001-01-01 to 9999-12-31 • recommended over legacy `DATETIME`.
DATE	`birth_date DATE`	• Date-only type (no time component) storing YYYY-MM-DD • 3 bytes storage.
TIME	`event_time TIME`	• Time-only type (no date) with 100-nanosecond precision • range 00:00:00.0000000 to 23:59:59.9999999.
BIT	`is_active BIT`	• Boolean-like type storing 0, 1, or NULL • multiple BIT columns in one table share storage bytes.
UNIQUEIDENTIFIER	`session_id UNIQUEIDENTIFIER`	• 16-byte GUID for globally unique values • typically populated via `NEWID()` or `NEWSEQUENTIALID()`.
JSON	`metadata JSON`	• Native JSON type (new in 2025) with optimized binary storage up to 2GB • supports indexing and faster access than `NVARCHAR`.
VECTOR	`embedding VECTOR(768)`	• Vector data type (new in 2025) for AI embeddings and similarity search • stores fixed-dimension floats in single-precision (4-byte) or half-precision (2-byte).
VARBINARY(n)	`file_data VARBINARY(MAX)`	• Variable-length binary up to `n` bytes • `MAX` allows up to 2GB for storing files, images, or encrypted data.
XML	`config XML`	• XML document storage with schema validation and XQuery support • max 2GB.

Table 2: Query Fundamentals

These are the clauses every SELECT is built from — filtering rows, sorting them, grouping for aggregation, and limiting how many come back. Master the order they execute in and you can read or write almost any query with confidence.

Clause	Example	Description
SELECT	`SELECT id, name FROM users`	• Retrieves specific columns from a table • use `*` for all columns (avoid in production for performance).

The rest of this section is part of the full cheat sheet

Table 3: JOIN Operations

Joins are how you stitch related tables back together, and picking the right one comes down to which unmatched rows you want to keep. Beyond the standard inner and outer joins, the APPLY operators let you run a correlated subquery or table-valued function once per outer row — a trick that solves problems plain joins can't.

Type	Example	Description
INNER JOIN	`FROM orders o` `INNER JOIN customers c ON o.customer_id = c.id`	• Returns only matching rows from both tables • most common join type for related data.

The rest of this section is part of the full cheat sheet

Table 4: Aggregate Functions

Aggregates collapse many rows into a single summary value — counts, totals, averages, and extremes — and almost always pair with GROUP BY. SQL Server 2025 widens the toolkit here with a multiplicative PRODUCT and the JSON_OBJECTAGG/JSON_ARRAYAGG builders that roll grouped rows straight into JSON.

Function	Example	Description
COUNT	`SELECT COUNT(*) FROM orders`	• Returns total row count • `COUNT(*)` includes NULLs, `COUNT(column)` excludes NULLs.

The rest of this section is part of the full cheat sheet

Table 1: Core Data Types

Type	Example	Description
INT	`product_id INT`	• Standard 4-byte integer ranging from -2,147,483,648 to 2,147,483,647 • the most common whole number type.
BIGINT	`order_total BIGINT`	8-byte integer supporting extremely large values (-9,223,372,036,854,775,808 to 9,223,372,036,854,775,807).
VARCHAR(n)	`name VARCHAR(100)`	• Variable-length string storing up to `n` characters (max 8,000) • uses only space needed plus 2 bytes overhead.
NVARCHAR(n)	`title NVARCHAR(200)`	• Unicode variable-length string supporting international characters • 2 bytes per character, max 4,000 characters.
DECIMAL(p,s)	`price DECIMAL(10,2)`	• Fixed-precision numeric where `p` is total digits (1-38) and `s` is decimal places • essential for financial data.
DATETIME2	`created_at DATETIME2`	• High-precision datetime with 100-nanosecond accuracy • range 0001-01-01 to 9999-12-31 • recommended over legacy `DATETIME`.
DATE	`birth_date DATE`	• Date-only type (no time component) storing YYYY-MM-DD • 3 bytes storage.
TIME	`event_time TIME`	• Time-only type (no date) with 100-nanosecond precision • range 00:00:00.0000000 to 23:59:59.9999999.
BIT	`is_active BIT`	• Boolean-like type storing 0, 1, or NULL • multiple BIT columns in one table share storage bytes.
UNIQUEIDENTIFIER	`session_id UNIQUEIDENTIFIER`	• 16-byte GUID for globally unique values • typically populated via `NEWID()` or `NEWSEQUENTIALID()`.
JSON	`metadata JSON`	• Native JSON type (new in 2025) with optimized binary storage up to 2GB • supports indexing and faster access than `NVARCHAR`.
VECTOR	`embedding VECTOR(768)`	• Vector data type (new in 2025) for AI embeddings and similarity search • stores fixed-dimension floats in single-precision (4-byte) or half-precision (2-byte).
VARBINARY(n)	`file_data VARBINARY(MAX)`	• Variable-length binary up to `n` bytes • `MAX` allows up to 2GB for storing files, images, or encrypted data.
XML	`config XML`	• XML document storage with schema validation and XQuery support • max 2GB.

Table 2: Query Fundamentals

Clause	Example	Description
SELECT	`SELECT id, name FROM users`	• Retrieves specific columns from a table • use `*` for all columns (avoid in production for performance).

Table 3: JOIN Operations

Type	Example	Description
INNER JOIN	`FROM orders o` `INNER JOIN customers c ON o.customer_id = c.id`	• Returns only matching rows from both tables • most common join type for related data.

Table 4: Aggregate Functions

Function	Example	Description
COUNT	`SELECT COUNT(*) FROM orders`	• Returns total row count • `COUNT(*)` includes NULLs, `COUNT(column)` excludes NULLs.