Overview of Public Datasets

Users with folder management permissions in Public Data can add data tables to folders.

These data tables can be directly referenced by other users with use permissions.

Four types of datasets are available in Public Data: database tables, SQL datasets, Excel datasets, and self-service datasets.

These four types of datasets can be further divided into two categories: basic tables and self-service datasets.

• Basic table: table which can be directly added to Public Data.

• Self-service dataset: dataset formed from edited basic tables or Excel data (for details, see [Analysis Subject] Data Overview) which you can apply for publishing to Public Data.

• Fields identified as date types by BI will automatically be completed as yyyy-MM-dd HH:mm:ss (example: 2020-10-15 00:00:00).

• Decimals (if any) in fields will definitely be identified as numeric fields.

• When no decimal exists in fields:

If Data Type Identification is disabled, values less than or equal to 15 digits are identified as numeric fields, and values greater than 15 digits are identified as text fields.

If Data Type Identification is enabled, all values are identified as numeric fields no matter how many digits there are. The fields are identified as double fields with effective precision to be 15 bits (2^52). Therefore, fields with over 15 bits may lead to precision loss.

• For tables added without parameters enabled (direct-connected and distributed):

If these tables have not been edited, numeric fields are obtained after you enable parameter restart and enter the table edit page.

If these tables have not been edited (field type conversed in 5.1.5 and later versions), text fields remains unchanged (not read as numeric fields) after you enable parameter restart and enter the table edit page.

Data Type Identification is disabled by default. To enable it, see BI/Spider Parameter.