Now a variety of report styles, more and more cool so people could attend. If you want to be traceable from the data results, because the data processing deep-rooted and too technical, business people make ends meet. How to make business people see a clear visualization of data from the source, the operation step by step process to show, Wing Hung science and technology to help you out.
Today article, the main referral Shao Yonghong self-service data query, to provide a powerful and convenient data preparation and integration, the user may need only a simple drag and drop operation and visualization on the graphical interface through this way, you can Construction of complex data sets.
Data preparation process involves three main functional nodes: an input node (connected to the database table for realizing the import EXCEL data, creating line data), the intermediate node (input node and associated data manipulation node) node and the query result.
First, the input node
I.e., data input node of the source node entry. You can connect to the database tables, import Excel, create an embedded data mode, query data from different types of data as input node self-service operation.
Second, the association switching nodes &
Operating system embedded in the ribbon association (connection joint) and the conversion (reverse pivot tables, grouped and summarized column from the circulation, the mirror, to weight) two types of data processing node.
1, associated node - coupling (the Join)
Between the table and the data table associated with the established connection by connection type nodes, data integration. If the information defines an outer join between database tables, joins node will do based on these information will default outer join join.
2, associated node - United (Union All)
By way of the connection, the combined node may be the same number of columns in the data table (including duplicate rows) stitched together.
Example: A table includes transaction information on the 1st to 20th, B table includes the transaction information 21 to 30 days, is to integrate two table data together by a joint node, it is possible to view a full cycle of the 1st to 30th Trading Information.
3, the conversion - an inverse PivotTable (Unpivot Table)
Unpivot node is multidimensional tables into a one-dimensional data processing operation.
Example: the figure below have the same user mathematics, language, English three different subjects at Grade, now need to be integrated into all subjects form a single row in the new data set by data processing, facilitate each student's summary of the results analysis.
在逆透视节点中，设置数据保持不变列为保留列和需要从多维转一维的列为转换列。
即可应用该数据集按班级、人员进行各科成绩的汇总分析，转换结果如下图所示。
4、分组和汇总 ( Aggregate)
分组和汇总即对数据进行分组聚合。
举例：在上例的基础上，现对每位同学的成绩进行分组，并汇总每位同学的总成绩。
5、自循环列 ( Autoloop)
自循环列就是根据设置自动分出层级关系的列，每个查询只能创建一个自循环列。
通常我们会遇到一列中数据存在层级关系，如河北省、石家庄、正定县，三个区域存在所属关系，在区域ID上也存在父子级关系。通过自循环列操作可实现各个区域的层级划分。
通过自循环节点进行层级处理后结果如下图所示：
6、镜像
通过连线镜像节点，任意节点可以被复制一个或多个，在需要对同一输入节点并行进行多种数据处理的场景下应用。
7、去重
通过连线的方式，去重节点可以把连线节点的重复记录去掉。
举例：A表中用户1和用户数据2存在重复数据，现对其进行去重，原数据如下。
连接去重节点之后重复记录被去掉。
三、查询结果节点
查询结果节点是所有节点数据处理的终结点，查询结果节点只能有一个输入。可通过操作查询结果节点将自服务数据集导入数据库。
四、实时性能检测
自服务数据准备中，数据执行的绝对快慢很难得知，可以通过连线的颜色进行实时性能检测，连线颜色为绿色时表示性能快，连线为黄色时表示性能慢，用户可以根据性能快慢进行性能调优。
如图则性能一般，需要进一步调优。在调整后可以点击画布左上角性能检测，小球会沿着输入节点的轨迹，开始滚动，在查询结果节点停止，小球的颜色跟连线的颜色相同反应性能。
五、自动布局
自动布局功能为用户提供了节点和节点之间根据既定算法优化布局，让布局更加合理美观。
自动布局前：
自动布局后:
总结
In service query from the user through a visual workflow manner, it has been built directly query data model with mild conversion, but can also be introduced directly into the external variety of data sources and data conversion model. Not only provide self-service inquiry packet and data summary table conversion functions inverse perspective transformation of the data structure, while increasing the variety of new functions such as data conversion: value map, remove spaces, removing duplicates, replace missing values, split as multiple columns, create a group and so on.
Self-service queries the user can make the process smoother operation, user-friendly interface full view of the data structure and details of view to understand. By wire color and prompt, friendly and real-time display of performance data for each node, and display the data model of the overall performance of the state timely notice to the user.