Database Reference
In-Depth Information
tables of virtually any size.
Table 2.2
summarizes the expanded
functionality.
Table 2.2
Removing Limits
Limitation
Date
Changed
Description
JOIN
operations required one
table to be smaller than 8 MB.
March
2013
Now,
JOINs
of virtually
any size can be used via
the
EACH
keyword.
GROUP BY
operations that
resulted in more than a few
million distinct results could cause
out-of-resources errors.
March
2013
GROUP BY
operations of
virtually any size
supported via the
EACH
keyword.
Query results required to be
smaller than 128 MB.
June
2013
The allowLargeResults flag
was added to allow results
of virtually any size to be
returned.
Streaming inserts limited to 1000
rows per second.
March
2014
Limits were raised to
100,000 rows per second,
per table.
In addition, quota sizes have been raised considerably over time. When
launched, BigQuery supported only 2 concurrent queries, and 100 load jobs
per day that could import only 50 GB at a time. Now, BigQuery supports up
to 20 concurrent queries per project, and 1,000 load jobs per table per day.
Each load job can import up to 1 TB. If you're using the streaming interface
to add data, you can insert up to 100,000 rows per second per table.
Future Predictions
By the time you read this, some of the exciting new features that the
BigQuery team is working on now will be released. Google has a policy of not
speculating about new features before they happen. That said, we want to
give you a couple of hints about what you might expect from BigQuery in the
future. Rather than reveal particular features, we'll just describe what kinds
of things are important to us and that we'll be working on improving.
