News from the AI & ML world
DeeperML - #bigquery
|
@github.com
//
Google is enhancing SQL with a new pipe syntax in BigQuery and Cloud Logging. This extension to GoogleSQL introduces a linear, top-down approach to writing queries, using the pipe operator to chain operations such as filtering, aggregating, and joining. This method simplifies data transformations and makes queries more intuitive, readable, and maintainable, contrasting with the rigid clause structures of traditional SQL which often require complex subqueries. Since its general availability in April, users have embraced pipe syntax for streamlining data transformations, building insightful reports, and efficiently analyzing logs.
The key benefit of pipe syntax lies in its ability to create a more natural flow of data transformations. Operations can be applied in any order, reducing the need for cumbersome subqueries or Common Table Expressions (CTEs). For example, building a recommendation report to determine the most cost-effective pricing model for long-term data storage in BigQuery can be simplified using pipe syntax compared to the more verbose standard SQL. This improved readability and flexibility allow developers and data analysts to work more efficiently. Beyond SQL enhancements, Google is also advancing healthcare through AI with the release of open MedGemma AI models. Unlike previous models locked behind expensive APIs, these tools are made available to healthcare developers, allowing them to be downloaded, modified, and run as needed. MedGemma models, including the flagship MedGemma 27B Multimodal, can process medical text and analyze medical images such as X-rays and pathology slides. These models have shown impressive performance on medical knowledge benchmarks, offering a cost-effective solution for healthcare systems and potentially transforming patient care.
Share:
References :
Classification: |
BenchmarksBlogsResearch Tools |