Generate sophisticated query for Google BigQuery in simple way.
This project is archived and no longer maintained.
BQX is a minimal query generator for Google BigQuery. It's mainly intended for being used by data analysts / scientists who want to analyze big data.
BQX fires its power especially on making LONG and NESTED query. BigQuery doesn't accept two or more queries at once so user has to make views or make terribly nested query. BQX's features help you make long query preserving high readability. See example for its features and usage.
- Python 3.4 ~
pip install bqx
>>> from bqx.query import Query as Q
>>> from bqx.parts import Table as T, Column as C
>>> from bqx.func import SUM
>>>
>>> shakespeare = T('publicdata:samples.shakespeare')
>>> word = C('word')
>>> q = Q().SELECT(word).FROM(shakespeare).LIMIT(10)
>>> print(q.getq())
SELECT word
FROM publicdata:samples.shakespeare
LIMIT 10
See example directory for more detailed examples.
This library is BSD-licensed.
Running example on REPL is good for getting started, and we suggest you to use BQX with front-ends like pandas for next step.
It aims to generate query excecuted on BigQuery but it might be applied to other SQL environments because difference between plain SQL and BigQuery is small.
We are using some compute engines which can process SQL (and dialects) like BigQuery, Hadoop and Spark. So in near future, adding other SQL dialect is planned.
The first reason is for avoiding collision with Python's reserved words. The another reason is for SQL's UPPERCASE manners.
We hope it's so. All main clauses are implemented from v0.2.0, but not fully tested by hand yet.
We're sorry but documentation is not available right now. It will be published on ReadTheDocs.org soon. Stay tuned.