User guide-MaxCompute(MaxCompute)-阿里云帮助中心

This topic recommends different learning paths based on your role.

If you are a MaxCompute beginner

If you are a beginner, familiarize yourself with the following modules first. You can then study other modules in more detail.

Module	Description
Product Introduction	Outlines the overview, main features, scenarios, limits, and basic concepts of MaxCompute. Reading this chapter gives you a general understanding of MaxCompute.
Preparations	Guides you through preparing an account, setting up an environment, creating a table, importing data, running SQL, and exporting result data.
Quick Start
Common commands	Lists common MaxCompute commands. This helps you become more familiar with operating MaxCompute.
Tools	Master the tools used with MaxCompute, such as the client and MaxCompute Studio, before you analyze data.
Endpoint	Describes the connection methods and Endpoint information supported in each MaxCompute region. It also explains network connectivity and data download fees when you access other cloud products such as ECS, Tablestore, or OSS.

If you are a data analyst

If you are a data analyst, read the SQL module to learn how to query and analyze large-scale data stored in MaxCompute. MaxCompute SQL supports the following main features.

Feature	Description
DDL operations	Supports managing tables, partitions, columns, lifecycles, and views.
DML operations	Supports inserting or updating data in tables and partitions.
DQL operations	Supports various query operations such as SELECT and subqueries.
Enhanced operations	Supports enhanced SQL operations, such as using commands to import and export data from MaxCompute tables and replicate table data.
Built-in functions	Supports data processing using built-in functions such as mathematical functions, window functions, date functions, aggregate functions, and string functions.
UDF	Supports creating user-defined functions to meet more computing needs.

If you have development experience

If you have development experience, an understanding of distributed concepts, and want to solve data analytics problems that cannot be handled by SQL, you can learn about the advanced feature modules of MaxCompute.

Module	Description
MapReduce	MaxCompute provides a Java MapReduce programming model. You can use the interfaces (Java API) provided by MapReduce to write MapReduce programs and process data in MaxCompute.
Graph	An iteration-oriented graph computing framework. It uses graphs for modeling. A graph consists of vertices and edges. Vertices and edges contain values. The graph is edited and evolved through iterations to produce the final result.
Tunnel	You can use the Tunnel service to upload offline data to MaxCompute in batches or download offline data from MaxCompute.
Java SDK	Java interfaces provided for developers.
Python SDK	Python interfaces provided for developers.

If you are a project Owner or administrator

If you are a project Owner (who creates and uses projects) or an administrator (who manages projects, security, and costs), you should familiarize yourself with the following modules.

Module	Submodule	Description
Project management	Preparations for creating a project	A project is the basic organizational unit in MaxCompute. It is similar to a database or schema in a traditional database. A project is the main boundary for multi-user isolation and access control. A user can have permissions on multiple projects at the same time. Through security authorization, a user can access objects in another project, such as tables, resources, functions, and instances. Using MaxCompute involves operating various objects within a project. The preparations are as follows: Resource budget Billable MaxCompute resources mainly include storage, computing, and Internet download traffic. Storage resources: Billed on a tiered pay-as-you-go basis. You can estimate the cost based on the data volume using a formula. The budget is not an absolute value because data is not stored in MaxCompute all at once, and data is imported and exported at all times. Computing resources: Computing resources are available in pay-as-you-go and subscription modes. It is difficult to estimate the usage of computing resources initially. Use the pay-as-you-go mode first. After a period of testing, decide whether to switch to the subscription mode based on your usage. Outbound traffic over the internet: Billed on a pay-as-you-go basis. You are charged only for downloads over the internet. For more information about billing, see Billable items and billing methods. Prepare an account and activate the service Before creating a MaxCompute project, you must activate the MaxCompute service. Only an Alibaba Cloud account can be used as the root account, and this account is the billing entity. After you confirm the account, select the pay-as-you-go or subscription mode when you activate the MaxCompute service based on your resource budget.
	Create a project	For more information about how to create a project, see Create a MaxCompute project. When you create a project, consider whether to select a standard mode or basic mode project from a business perspective. From a security perspective, consider whether to use a personal account or a compute engine-specified account. For more information, see MaxCompute access identities and MaxCompute data security management guide.
	Project member management	Member management mainly involves member responsibilities and security issues. If you use MaxCompute through DataWorks, you need to consider the permissions required for both products. For more information, see Permissions for MaxCompute and DataWorks.
	RAM user management	MaxCompute projects support two account systems: Alibaba Cloud accounts and Resource Access Management (RAM) users. You can add any RAM user under your Alibaba Cloud account to a MaxCompute project. For more information about RAM users, see Create a RAM user. When you use MaxCompute and DataWorks workspaces through DataWorks, you can only add RAM users under an Alibaba Cloud account as members. Therefore, the Alibaba Cloud account must create and maintain RAM users through the RAM system. Note MaxCompute does not define permissions in RAM. Therefore, the permissions for RAM users must be configured using MaxCompute commands or the project management feature of DataWorks. For more information, see MaxCompute data security management guide. Assign one RAM user to each project member. Do not let multiple members share one RAM user. For members who have left the company or changed positions, promptly clear their RAM user accounts. If a RAM user is added as a project member in DataWorks, first remove the project member and then delete the RAM user in the RAM system.
	Scheduling resource management	Scheduling resources Schedule resources in DataWorks execute tasks dispatched by the scheduling system. These resources are available in the following two modes. For more information, see View the resource group list. Default scheduling resources. These are the public resource pools of DataWorks. When the number of concurrent nodes in DataWorks is high and scheduling resources are tight, tasks enter the 'waiting for scheduling' state. The nodes start to execute the issued tasks only after they obtain the resources. Custom scheduling resources. You can configure the ECS instances that you purchase as scheduling servers to execute and distribute tasks. An Alibaba Cloud account can create custom scheduling resources. These resources include several physical servers or ECS instances and are mainly used to execute data synchronization or other tasks. Exclusive resources DataWorks provides exclusive scheduling resources, exclusive data integration resources, and exclusive resources for running DataStudio in the production environment for you to choose from to ensure optimal performance. For more information, see DataWorks resource groups.
	Project settings	During project development, some project settings must be configured by the project Owner. For example, setting whether to allow full table scans for a project and setting whether to enable data type V2.0 by default for a project. For more information, see Project operations.
Security management	Personnel management	Security management includes personnel management, role management, and permission management. When you use MaxCompute through DataWorks, you need to understand the permission relationships between the two products because they have different permission models. Then, you can manage permissions as needed. During security management, you need to know how to grant permissions to users, share resources across projects, configure data protection for projects, and perform policy authorization: Permission management and List of security parameters. Security management examples: Provides security examples based on real customer business requirements.
	Role management
	Permission management
Cost management	None	Resource budgeting is used to estimate costs before use. Based on the billing methods of MaxCompute, many businesses cannot accurately estimate costs. Therefore, you need to manage costs throughout the business development process. Pay attention to the following items: Product billing and pricing. For more information, see Billing methods. Product overdue payment alerts and service suspension policy. For more information, see Overdue payment alerts and service suspension policy. Subscription upgrade and downgrade operations. For more information, see Upgrades and downgrades. Subscription renewal. For more information, see Renewal management. The product supports switching between pay-as-you-go and subscription billing methods. For more information, see Switch billing methods. View MaxCompute billing information. For more information, see View bills and Analyze the usage details of a MaxCompute bill.