Database Reference
In-Depth Information
Job Configuration
The
job configuration
section specifies what should get run. BigQuery may
tweak the configuration—it might canonicalize path names, for
example—but after the job has been created (that is,
Jobs.insert()
has
returned successfully) the configuration will never be changed.
There are four types of jobs: Query, Load, Copy, and Extract. Every query
that you run is a
Query job
.
Load jobs
import data from outside of BigQuery.
Copy jobs
make fast copies of tables.
Extract jobs
can be used to make entire
tables available outside of BigQuery.
The job configuration has a subsection for each type of job that can be
run. The presence of the particular subsection is the signal for BigQuery to
run that particular type of job. Only one per-job-type configuration section
should be present at a time. For example, a Query job configuration may
look like this:
{"query": {"query": "SELECT 17"}}
Alternately, a Load job configuration may look like this:
{"load": {
"sourceUri": "gs://foo/bar.csv",
"destinationTable": {
"projectId": "bigquery-e2e",
"datasetId": "logs",
"tableId": "latest"}}}
Don't worry about the exact fields that are present here. Individual job types
are discussed in more detail in subsequent chapters: Query jobs in Chapter
7 (“Running Queries”), Load jobs in Chapter 6, Copy jobs in Chapter 10
(“Advanced Queries”), and Export jobs in Chapter 11 (“Managing Data
Stored in BigQuery”). For now, we will limit discussion to a few settings
that are common across job types:
dryRun
,
createDisposition
, and
writeDisposition
.
Dry Run
The
dryRun
flag is one of the only settings present on the top level of the
job configuration, and its purpose is to instruct BigQuery to not actually run
