Your information might be saved within the geographic location that you simply selected on the dataset’s creation time. After a dataset has been created, the placement cannot be modified. One necessary consideration is that you simply will be unable to question throughout a number of places, you possibly can learn particulars on location concerns right here. Many customers selected to retailer their information in a multi-region location, nonetheless some selected to set a selected area that’s near on-premise databases or ETL jobs.
Entry to information inside BigQuery might be managed at completely different ranges within the useful resource mannequin, together with the Undertaking, Dataset, Desk and even column. Nevertheless, it’s typically simpler to manage entry increased within the hierarchy for less complicated administration.
Examples of frequent BigQuery undertaking buildings:
By now you in all probability understand that deciding on a Undertaking construction can have a giant affect on information governance, billing and even question effectivity. Many shoppers selected to deploy some notion of information lakes and information marts by leveraging completely different Undertaking hierarchies. That is primarily a results of low cost information storage, extra superior SQL choices which permit for ELT workloads and in-database transformations, plus the separation of storage and compute within BigQuery
Central information lake, division information marts
With this construction, there’s a frequent undertaking that shops uncooked information in BigQuery (Unified Storage undertaking), additionally known as a Knowledge Lake. It’s frequent for a centralized information platform workforce to create a pipeline that truly ingest information from varied sources into BigQuery inside this undertaking. Every division or workforce would then have their very own datamart initiatives (e.g. Division A Compute) the place they will question the info, save outcomes and create mixture views.