Running Queries - Google BigQuery Analytics - page 266

of your choosing. If you want to make a copy of the table, you can run a
Copy job. Both these options are available from the command line or
the web UI.
There are valid use cases in which you still want to use SELECT * . For
example, maybe you want to filter your table and you want to create a
smaller temporary table (as in SELECT * FROM . . . WHERE . .
. ). Using SELECT * in an inner query (as in SELECT foo FROM
(SELECT * from tablename) , which will read only the foo
column) is efficient because only the columns that are used in the outer
query actually need to be read.
The bottom line is this: If you use SELECT * in your query, you should
pause a moment and consider whether there is a better option.

Determining Query Cost
You can tell how much a query costs by reading the totalBytesProcesed
field in the response to Jobs.query() or in the statistics field in the
Job resource returned by Jobs.get() . Consider the following query, which
counts the number of active ZIP codes with nonzero populations in each
state:

>>> query = """
… SELECT state, COUNT(*) AS cnt
… FROM [bigquery-e2e:reference.zip_codes]
… WHERE population > 0 AND decommissioned = false
… GROUP BY state, ORDER BY cnt DESC
… """

This query reads three fields from the table
bigquery-e2e:reference.zip_codes : state , decommissioned ,
and population . Here is a Python command that runs the query and
returns only the number of bytes processed:

>>> service.jobs().query(
… projectId=project_id ,

… body={'query': query, 'useQueryCache': False}