The second step in the core process of the product is to create a data set and perform data modeling. If data is the sea, then creating a data set and modeling reasonably is the oars of rowing boats. Only with oar can swim in the ocean of data. Currently, there are two ways to create a dataset: (1) Create a dataset directly from the data table (2) Create a dataset through custom SQL to achieve preliminary modeling.
Create a dataset:
(1) Directly create a new dataset:
As shown in the figure below, on the data source list page, find the target table, click the Create Data Set button and select the storage folder to create a new one.
(2) Customize sql to create a dataset:
Currently, the product supports MaxCompute and MySQL data sources to create data sets with custom sql, which can realize multi-table association, add target fields, etc. As shown below, click 'Custom SQL' to enter the editing interface:
At present, there are some rules for custom sql to pay attention to:
MySQL
List of functions not supported:
MaxCompute
The MaxCompute data source natively supports functions.
common problem
Q: Why can't the target table be searched for the MaxCompute data source?
A: (1) Meta synchronization takes about 30 minutes to create a new table in the MaxCompute data source. Please wait a moment to ensure that the user AccessID and AccessKey of the synchronization table are valid.
Q: In the MaxCompute data source, when I create a table, it prompts that I do not have permission?
A: To apply for permissions, ensure that you have list, select, and create instance permissions.
Q: The MaxCompute data source data set editing interface refreshes the preview data and reports an error:
A: The reason for the error: The newly created project (project) at 20:00 on January 10, 2018 prohibits the full table scan of the partition table by default, that is, when the full table scan statement is executed for the partition table of the project, a failure will be returned.
1. Specify the partition conditions, and add 'where pt = xxx' after the SQL statement for the full table scan of the partition table to limit the partition.
2. Add a filter to the partition field when creating a report.
Q: SqlServer, PostgreSQL, Oracle data source, can not find the corresponding table?
A: Please confirm whether the schema added to the data source is correct
Q: To create a dataset with custom sql, the fields are in the form of the image below, and an error is reported when refreshing the preview data?
A: As shown in the picture above, what is the character type? Mark it, indicating that the field alias has been modified later. An underscore _ before and after the field name indicates that single quotes are added to the field alias. Solution: Modify the custom sql and save it, synchronize the table structure in the data set editing interface, delete the error field, save the data set, and finally refresh the preview data.
If you have any questions, please scan the QR code in the picture below to add the nail group, @元世青 for technical support!
