Database Reference
In-Depth Information
+-------------+--------------------+
| word | rel_freq |
+-------------+--------------------+
| holp | 207.0076923076923 |
| pricket | 79.81512605042016 |
| unpeople | 57.56363636363636 |
| burgonet | 55.54385964912281 |
| therewithal | 55.435797665369655 |
| nonino | 50.656 |
| extemporal | 46.10679611650485 |
| hungerly | 45.66346153846154 |
| prains | 42.78378378378378 |
| vizards | 40.851612903225806 |
+-------------+--------------------+
According to the Internet, “holp” is an archaic form of the past tense of
“to help,” as in “My colleague was hungerly trying to learn BigQuery, so I
donned my burgonet, and thus equipped with my vizards, by extemporal
query editing, I
holp
him therewithal.” This sentence would probably make
more sense to someone living in Elizabethan England.
Differences from Standard SQL
There are a number of differences between BigQuery SQL and standard
SQL. The reasons for these differences primarily fall into three buckets:
things that were more difficult to implement in a parallel query engine (for
example, an exact
COUNT DISTINCT
), syntax shortcuts for expected usage
patterns (for example, comma as
UNION ALL
rather than
JOIN
), and extra
features that aren't available in standard SQL (for example, handling of
nested and repeated fields). Some of the principal differences from standard
SQL are listed in this section.
Comma as UNION ALL
This is perhaps the most confusing quirk of BigQuery for people with a
background in SQL. In most variants of SQL, if you specify “…
FROM
table1, table2
,” this is actually an implicit
JOIN
operation. However, in
BigQuery, the previous operation is
UNION ALL
; that is, BigQuery performs
the query as if there was one large table with all the data from both
table1
and
table2
.
