BigQuery: All content tagged as BigQuery in NoSQL databases and polyglot persistence
New features added to Google BigQuery:
- Big JOIN: use SQL-like queries to join very large datasets at interactive speeds
- Big Group Aggregations: perform groupings on large numbers of distinct values
- Timestamp: native support for importing and querying Timestamp data
I read with interest both the announcement and the technical (?) details post about the new SQL keyword
EACH introduced by BigQuery to perform
GROUP BY for “large tables”. Unfortunately I couldn’t find what’s behind this new keyword.
This made me think again of what’s wrong with SQL: almost every engine implementation detail bubbles up to the user creating a new flavor of SQL. Just think about it:
EACH has no meaning for either of these operations; is there a
NOTEACH JOIN?. But it was needed to instruct the engine to perform the operation differently.
Original title and link: Goolge BigQuery: JOIN and GROUPBY EACH. And Something Is Wrong With SQL ( ©myNoSQL)
Announced at GigaOm Structure Data event, Google launches a new BigData service named BigQuery:
BigQuery enables businesses and developers to gain real-time business insights from massive amounts of data without any upfront hardware or software investments.
A quick bullet point list of BigQuery features and limitations:
- BigQuery is ideal for running queries over vast amounts of data—up to billions of rows—with great speed.
- BigQuery is good for analyzing vast quantities of data quickly, but not for modifying it. In data analysis terms, BigQuery is an OLAP (online analytical processing) system.
- You can import data into BigQuery as CSV data, where it is stored in the cloud in a relatively small number of tables with no explicit relationship to each other.
- BigQuery isn’t a database system:
- It doesn’t support table indexes or other database management features.
- BigQuery supports a specialized subset of SQL; it doesn’t support update or delete requests.
- BigQuery supports joins only when one side of the join is much smaller than the other.
- BigQuery can be used by any client able to send REST commands over the Internet.
After the break you can watch the 15 minutes video recorded at the GigaOm event.