Dataset

A dataset is a data model composed of one or more tables. It serves as the foundation for visual queries, charts, and dashboards. Before using these three modules, a data connection must be established in the "Dataset" module to import data into DataWind.

Data type

Data types are used to divide data into groups that require different amounts of memory. Large memory is only needed when working with large datasets. This enables optimal use of memory.

Expression

Expressions can be written using SQL to obtain the desired data. For example, if the original table contains a timestamp field create_time: 1522920936, but a month field is needed, a function can be used to extract a new field creat_month: 2019-08. The function is toMonth(toDateTime(create_time)).

Data lifecycle

The data lifecycle allows for the regular cleanup of unnecessary data based on dates, freeing up shared resources.

Synchronization type

Data from the specified database is periodically transferred to the DataWind database.
Scheduled: Data from the upstream database is produced and synchronized according to the specified time.
Paused: Synchronization stops after the initial synchronization is completed.

Primary key

Setting a primary key for a dataset improves query speed. Typically, the filter field specified in each query is selected as the primary key. There can be multiple primary keys. The first primary key field has the greatest effect. The importance of each subsequent primary key field decreases.

Partition

Datasets use partitions to horizontally distribute load, physically moving data closer to the users who access it most frequently.

Shard

Datasets use a sharding field to ensure data is evenly distributed across all nodes. Sharding can improve query speed.

Sampling field

When querying a dataset, sampling a small amount of data can improve query speed. By default, HASH is performed based on the sampling field.

Sampling ratio

When performing visual queries, sampling is conducted based on the sampling field. The smaller the sampling ratio, the smaller the data volume, and the faster the visual query.

Data modeling

The process of cleaning, processing, and producing metadata following specific data processing logic is called data modeling.

Data connection

The capability to configure connections with various data sources to enable data reading as the primary function.

Canvas

A visual representation of a data processing workflow, where each functional module is arranged in a directed workflow.

Operator

Each block on the canvas represents an integrated capability for data reading, processing, algorithms, and data output.

Connection

The directed relationship between operators, which refers to the data flow channel between two operators.

Dataset

See Dataset Overview, which is the basic modeling tool for data preparation, providing fundamental multi-table association data modeling capabilities. It also serves as a data container that can be used in products, including ClickHouse datasets and Hive datasets.

Real-time task

A visual modeling task where the data source is updated in real time and the data stream is processed in real time.

Offline task

A visual modeling task where the data source is updated on a daily or hourly basis (that is, updated once per day or per hour), and the data processing workflow is executed on a schedule or manually.

Tab

A tab is a collection of multiple pages, allowing different titles to be displayed for different pages.

Global filter

A filter that can be applied to multiple charts simultaneously; filters are used to narrow the range of data displayed in charts.

Layer order

When different elements overlap, they are displayed according to the layer order.

Group

Treats different elements as a whole, allowing them to be configured and moved together.

Markdown

Markdown is a plain text markup language. By using simple markup syntax, ordinary text content can be formatted with specific formatting.

Undo

Return to the state before the most recent operation.

Redo

Reverse undo.

Align left

Align two or more elements to the left.

Align vertical center

Align the center lines of two or more elements vertically.

Align right

Align two or more elements to the right.

Align top

Align two or more elements to the top.

Align horizontal center

Align the center lines of two or more elements horizontally.

Align bottom

Align two or more elements to the bottom.

Distribute horizontally

Distribute the center points of two or more elements evenly in the horizontal direction.

Distribute vertically

Distribute the center points of two or more elements evenly in the vertical direction.

Equal spacing guide

When changing the position or size of elements, a guide line appears when the spacing between the moving element and the reference element is equal, or when the same width or height is detected.

Guide line

When changing the position or size of elements, a guide line appears to assist with layout positioning.

Zoom

Zoom in or out on the page.

Page settings

Select from common page sizes or customize the page width and height.

Appearance settings

You can set the page background color or add grid lines to the page.

Chart settings

You can set whether to hide the chart title and configure the chart title format. For tables, you can configure display settings and choose between standard or adaptive table width and height.

Transformation settings

You can set the XY coordinates of elements, as well as configure the width and height of the element area.

Add description

Add explanatory notes to the chart.

Table

A table is the basic form for displaying data. By simply configuring dimensions and metrics, you can display the corresponding detailed data. Tables also support style configuration, conditional formatting, column sorting, fixed columns, and more.

Pivot table

A pivot table summarizes and displays data according to column and row dimensions. A pivot table can be displayed by simply configuring the column dimensions, row dimensions, and metrics.

Details table

A details table is a table used to display detailed data.

Bar chart

A bar chart is typically used to describe changes in data over a period of time or across different regions.

Horizontal bar chart

A horizontal bar chart can be regarded as a bar chart oriented horizontally and is usually used to represent changes in data over different times or under different conditions.

Line chart

A line chart is generally used to observe data trends.

Area chart

An area chart is similar to a line chart, except that the area between the line and the X-axis is filled with color. Suitable for describing data trends and the contribution of parts to the whole.

Dual-axis chart

A dual-axis chart refers to a chart with two Y-axes, often combining a bar chart and a line chart, providing a more intuitive display. Applicable for analyzing two data sets with large differences, different data trends, or for year-over-year and month-over-month comparisons.

Pie chart

In a pie chart, the central angle of each sector is determined by the proportion of the represented quantity to the whole, and it is commonly used to show the proportion of partial data to the total.

Map

The map feature supports displaying differences in quantities across different regions.

Bidirectional bar chart

A bidirectional bar chart uses both positive and negative bars to show value comparisons between categories.

Combo chart

A combo chart is typically used to compare and display multiple metrics and their changes under the same dimension.

Scatter plot

A scatter plot, also known as an X-Y plot, displays all data as points on a Cartesian coordinate system to show the degree of correlation between variables, with the position of each point determined by the values of the variables.

Word cloud

A word cloud is a visual representation of text data, composed of words arranged in a cloud-like colorful graphic, used to display large amounts of text data; the importance of each word is shown by its font size.

Histogram

A histogram has a shape similar to a bar chart but conveys a completely different meaning. A histogram involves statistical concepts. First, the data must be grouped, and then the number of data elements in each group is counted. In a Cartesian coordinate system, the horizontal axis marks the endpoints of each group, and the vertical axis represents the frequency. The height of each rectangle corresponds to the frequency, and such a statistical chart is called a frequency distribution histogram.

Indicator card

An indicator card can be used to visually display the value of a specific indicator and its related indicators or trends.

Sorting

The sorting feature automatically sorts indicators according to specified rules.

Top N

Only displays the data for the top or bottom entries.

Total

Performs summary calculations on the data.

Percentage

Percentage calculation can quickly determine the percentage that each value occupies in a column of data.

Comparison

Comparative analysis refers to the data differences caused by other influencing factors (comparison conditions) under the same data standards (the same dimensional indicators).

Year-over-year and month-over-month

Quickly calculates year-over-year and month-over-month values for the data.

Indicator template

Allows flexible modification of expressions by filling in parameters.

Attribution analysis

Analyzes significant increases or decreases in indicators, automatically generates reports, and identifies influencing factors.

Sampling

Performs sampling operations on the dataset.

Routine

Based on SQL query results, visualizes the data, generates and saves charts, and then performs routine data updates.

Monitoring

Monitors relevant indicators in the chart.

Dynamic dimension/dynamic indicator

Enables simple and quick switching between multiple dimensions or indicators in the chart.

Drill up and drill down

Creates a hierarchical structure based on certain fields, allowing for quick drill up or drill down between levels within a visualization item.

Alias management

Adds new names to the contents of dimension fields.

Data panel

Includes dimensions, metrics, and filters. When fields are placed in this area, a query will be performed.

Dimension

A dimension is the perspective for analyzing a metric, and typically has only a small number of discrete values.

Metric

A metric measures data using different aggregation methods, and in most cases is a number. When defining a metric, you need to specify the scope of the data to be calculated and the calculation method. This is referred to as the 'statistical standard'.

Project

A project refers to an independent workspace within the system, typically divided by business or requirements. The resources and permission controls of different projects are independent of each other. You can create analysis projects based on business topics, business departments, and other subjects. Within an analysis project, team members can create visualization charts and dashboards. All main product features are contained within specific projects. Each project is logically isolated, and operations and settings within one project do not affect other projects.

Resource

Refers to dashboards, datasets, charts, and more within the system.

User group

A collection of users.

Row-level permission control

Controls permissions for row-level data in datasets.

Column-level permission control

Controls permissions for column-level data in datasets.

Workbook

A collection of SQL queries created for a specific need, containing multiple SQL query editors and query tasks.

Editor

SQL query code editor.

Query task

Refers to a single SQL query task.

Data source

Refers to the source of data for SQL queries. Currently, supported data sources for SQL queries are Hive and Clickhouse.

Cluster

A cluster is a group of service entities working together to provide a service platform with greater scalability and availability than a single service entity. Offers scalability and high availability.

Queue

Specifically, the cluster task execution queue.

Parsing

Performs syntax parsing for SQL query statements.

Ad hoc query

A query editor will be added after the [Temporary Workbook], rather than opening a new workbook each time.

Batch query / batch search

This means splitting SQL tasks into individual subqueries according to the batch query date, executing them separately, and then merging the data after the queries are completed.

System

System refers to the entire intelligent data insight platform. Whenever the term DataWind system, BI system, or system appears in the documentation, it refers to this concept. Typically, multiple projects may be created within your system.

Project

A project refers to an independent workspace within the system, generally divided according to business or requirements. Resources and permission controls are independent between different projects. You can create analysis projects based on business topics, business departments, and other themes. Within an analysis project, team members can create visual query charts, dashboards, and more. The main product features are all within specific projects. Each project is logically isolated, and operations and settings within one project do not affect other projects.

Multiple departments or multiple project teams