Skip to content

Latest commit

 

History

History
237 lines (164 loc) · 9.18 KB

widget.md

File metadata and controls

237 lines (164 loc) · 9.18 KB

Widget

** Widget -> new-> Datasource/Dataset (Drop-down Selection)**

Load Data

Load from existing dataset

If you load data from existing datasets, you can use the pre-defined Calculated Measures, Filter, schema tree and other features of Dataset. It's not allowed to create a new Calculated Measure on a chart design page when you use datasource create widget!

Load from direct from datasource

The default way to load data is to read from the existing dataset. You can switch load data type and write query script to experience Type SQL, Get Chart

Read data from Saiku2.x

Input the Repository of your Saiku as below

Saiku Repository CBoard Saiku Query
![](../../assets/Saiku_Rep.jpg) ![](../../assets/saikuquery.jpg)

Design Widget

![](../../assets/config_widget.png)

Widget Design Page

Page districts Introduction

Chart list Area

  • List saved charts
  • The upper-left function keys allow you to create, copy, edit or delete a chart. You need to select a single chart first, except creating a new chart.
  • Delete can only apply on a single chart
  • The directory tree structure is automatic generated by path name rule as the [Folder]?/[SubFolder]*/[chart name]
  • Folder is virtually based on file name
  • ** Double-click chart ** to edit
  • Select a chart and press Delete button on keyboard is short cut to delete a chart

Query Area

  • The query area defaults to select the existing datasets(Cube)
  • The bottom two function buttons are switch query mode (Switching the query between the dataset and the new AD-Hoc query) and Load data(JDBC offline data source data is loaded into the server cache, whereas the DataSource that supports and been setting to DataSource aggregation will only query for schemas)
  • Current operation status
    • New Status: every save action will create a new chart, which will facilitate to continuous design of multiple charts with different granularity and different presentation patterns.
    • **Attention **:If you do not need a chart name in a new state, clicking Save in succession will prompt you for a rename error
    • ** Edit (Edit)**Status: modify to save to affect the current report, you can click Save again

!> In New status, if you don't change the filename when click Save button again, you will see error info of duplicate name

Model area

  • The hierarchy tree in the model area needs predefine in dataset design page, otherwise you can only see a flat selects list
  • The model contains four parent nodes
    • Dimension
      • The columns below the hierarchy can be drilled / rolled under the cross table model
    • Measures
    • Calculated Measures: need Pre-defined, can only be used, can not be modified
    • Filter

Design area

  • Chart name format: [Folder]?/[SubFolder]*/[ChartName]
  • Switch chart type: different types of charts have different restrictions on input, hover on the chart type ICON to check restriction
  • Drag and Drop for OLAP and design
Design area target box Model area node type
Row box , Column dimension、Filter Dimension node
Value Box Measure node, Calculated Measure node
Filter Box Dimension node, Filter node
  • Function button
    • Preview
    • Preview query: view the dynamic query script for data source aggregation query
    • Save
    • Cancel: revert design change

Preview Area

  • Preview: Preview widget
  • Query: Debug query script
  • Option: Chart Setting

Fundamental★

In theory, multi-dimensional analysis is based on multi-dimensional data, multiple sequences series share the same index, the most intuitive representation of the data is the cross table

Cross table composition Chart design field
Row Header Nodes in a row
Column Header Nodes in Column Combined Measure in Value box
Aggregate data area Aggregate function apply on data set
  • As shown in the figure below, the header merge of the can only be correct when the header is sorted correctly

!> Please make sure you have get what's cross table and how to design a cross table before Go Ahead!

Slice and Spice

  • The node placed in the Row box and Column box can be filtered by clicking the edit button
  • You can also place dimensions in the Filter box to filter data purely

!> Make sure you are not loading a large size list or continuous value before load dimension members!

  • Filtering comparisons only support string and numeric comparison, because time functions differ from database to database, but most database datetime data type can be compared to a standard date string. Some databases may have index invalidation due to data type conversion. As of now, there is no special processing on datatime dimension here, please do some pre-processing on this piece while the data is ready, and debug with preview query at all times.
  • Columns on values can also be sorted and filtered
  • Support for input range comparison
  • TOP N display

!> As you can see from the form of the cross table, there is a sort conflict between the value sort and the row header sort, and the row sort has only one sort of value and only one is in force. **

Basic Graphics

Bar/Line Chart

One column for one line for a series bar.

Design District Chart requirement
Row Dimension x-axis Place one or more dimension nodes
Column Dimension category Place zero or more dimension nodes
Value Place one or more metric nodes (** Please don't ask us repeatedly why we don't have graphics when we don't have indicators **)
Addition axis Display double axis It is recommended that different show type axes be configured
Vertical / Horizontal X/Y axis position exchange -

Bar chart

One column for a pie chart

KPI card

KPI card needs no dimensional information, only one measure can be selected. formatted Formtter reference numbro api

Funnel chart

One row for a funnel.

Normally, a funnel needs to be displayed such as display -> click -> submit -> pay, a series of values for different measures, with multiple columns placed in the cross table for the value axis The values in the row are sorted automatically by size to form a funnel. ** So a row of funnels, a funnel. **, The following Demo makes no real sense, just as a demonstration

Sanky chart

Use row header and column header as nodes and value in data area as line weight. Crosstable can be considered as a joint-matrix for Sanky chart.

Why does my Sankey Diagram have no hierarchy?

A lot of people ask why there is no hierarchy in the Sankey Chart. The answer is that the hierarchy of the Sankey is related to your data itself.

Data include links `A->B and B->C`, then `B` will be a intermediate layer. Note also that ECharts requires data and cannot be looped \(A->B..->A\)

Radar

One column for one circle on Radar.

Bubble diagram

Design District chart requirement
Row dimension x-axis Place one or more dimension nodes
Column classify Place zero or more dimension nodes
Value Y-axis, bubble size, color depth One attribute per measure node

Contrast

Design District chart requirement
Row dimension Y-axis Only one dimension node can be placed
Value X-axis Only two index nodes can be placed

Word Cloud

The cartesian product is labeled according to multiple rows.

Tree Map

Design District chart requirement
Row dimension Multiple row dimensions represent multiple layers, which are classified by color Place one or more dimension nodes
Value The area is the number Only one index nodes can be placed

Heat Map

Design District chart requirement
Row dimension X-axis Place one or more dimension nodes
Column classify Place zero or more dimension nodes
Value Only one index nodes can be placed

Relation

Design District chart requirement
Row dimension center point sets Place one or two dimension nodes
Column classify Place one or two dimension nodes
Value Only one index nodes can be placed