Alibaba ACP Big Data Certification

Data+|Big Data Basic Services

MaxCompute

MaxCompute saves data to the system, the Column-Store file format is used to reduce invalid disk read operations (similar to Spark)
Project is the basic organizational unit of MaxCompute. It is similar to the concept of database or Scheme in traditional database, and is the main boundary of multi-user isolation and access control
The table is the data storage unit of MaxCompute.

MAC (Mandatory Access Control)

Column is the granularity of sensitive data can be controlled by Label Security

UDF & UDT & UDJ

UDF
- user defined scalar function
  
  Default UDF, if not specified, udf is udsf
  
  input and output is one-to-one relation
- user defined table valued function
  
  input and output is multi-to-multi
- user defined aggregation function
  
  input and output is multi-to-one
UDT

User defined type
UDJ

User defined join

E-MapReduce

Job

To create a job –> to create a configuration on how to run a job
Configuration includes:
- Jar packages to be run
- Input & output data addresses
- Running parameters
Job cannot be run dirrectly.

Execution plan

A bond that associates the job and the cluster.
multiple jobs can be combined into a job sequence and prepare a running cluster for the job (or automatically create a temporary cluster or associate an existing cluster)

DataWorks

DataWorks, Platform as a Service from Aliyun, based on MaxCompute supporting Data Integration, Data Development, Data Management, Data Quality Campaign, Data Sharing services.

Scheduling

Monthly scheduling: It supports multi days scheduling, while end of month is not supported, like 31th, 30th.
Minutes scheduling: The minimum scheduling cycle is 5 minutes, and it only supports 5*n minutes as scheduling interval.
It also supports weekly, daily, hourly scheduling.
Scheduling workflow is a DAG, current workflow will wait until its parent workflow finishes, even if sheduling time comes.

Data+|Big Data Analysis & Visualization

Data V

We want to add to the canvas from the widget menu on toolbar, many kinds of widget are currently supported by DataV, also third-party widgets, like E-Charts.

Quick BI

Workspace

Workspace is the place where team menber can collaborate. If number of team menber is over 10, Quick BI pro version, else the basic version.

In basic version, workspace is named personal workspace, while workspace in pro version.
Difference between personal workspace and workspace:
- A personal workspace is created automatically after the first logon. A workspace is created manually by an administrator of the organization.
- A personal workspace is unique to each Alibaba Cloud account or RAM user. It cannot be deleted.
- You are not allowed to add other members to a personal workspace. Therefore, a personal workspace does not support collaboration or sharing.
- For a personal workspace, data objects can be transferred to or shared with any Alibaba Cloud accounts. For a workspace, data objects can only be transferred to or shared with the members of the workspace.

Permission Management

Permission management include data object management and row-level permission management.
- Data objects include data sources, datasets, workbooks, dashboards, and portals.
Data object management is classified based on the workspace types, which include the personal workspace and the workspace.
Data object sharing
- Only the owner of the data object and the administrators of the workspaces have permissions to share the data object.

Details on ACP Exam