Overview
Near Real Time BI Connector provides controlled access to TI’s analytics warehouse, Redshift.
This direct access point provides the full breadth of tables used to power it’s award-winning analytics product, Reporting Hub.
Use Cases
NRT BI is built for analytics and business intelligence use cases, providing a seamless way to connect to a number of popular BI tools such as Power BI and others.
For use cases focused on data transportation to data lakes or warehouses the use of third party ETL tools is recommended.
Sync Frequency and Data Freshness
Every sync contains a "pipeline job" responsible for updating one or more tables, and runs every five to ten minutes. When that sync fires, it first checks to see if the past job for that pipeline has completed.
Common Columns
There are a number of columns that persist across many (sometimes all) tables, including:
Ids
Most primary and foreign key ids are uuid's but stored as strings(varchar 36).
In some cases ids will instead be keys from other systems and will have a different format. The docs for each table will provide details on ids if there is any specific usage.
Timestamps
All timestamps have the same format and structure unless stated otherwise.
- Format is
yyyy-mm-dd hh:mm:ss.dddddd
- eg,
2024-10-15 12:06:32.523754
- eg,
- Some timestamp columns may use milliseconds (3 decimals) instead of microseconds (6 decimals)
- Timestamps are not timezone aware and timezone is always UTC
- Max datetime is
9999-12-01 00:00:00.000000
- Values in the year of 9999 or greater will be pushed up or down to the max datetime as usage of the year 9999 is assumed that the user wants the max datetime
- eg
300000-05-26
will become9999-12-01
, but9999-02-15
will also become9999-12-01
- eg
Numeric Precision
If size of a number is not specific in the schema section for a table doc then its the size specified in the redshift datatype docs. https://docs.aws.amazon.com/redshift/latest/dg/r_Numeric_types201.html
created and updated at
These are timestamps that are populated for each record on every table. Every table will have different methods and processes for updatedat but this will always represent the last time a record was changed. Both createdat and updatedat will be standard TI timestamp format.