BigQueryIO - Must specify numFileShards when using FILE_LOADS with unbounded PCollection

Since Beam v2.2 it's possible to use [FILE_LOADS](https://beam.apache.org/documentation/sdks/javadoc/2.2.0/org/apache/beam/sdk/io/gcp/bigquery/BigQueryIO.Write.Method.html#FILE_LOADS). The documentation states that we have to specify _withTriggeringFrequency_ when using it, but doesn't talk about _withNumFileShards_, whereas if we don't specify it we get the below exception:

```

Exception in thread "main" java.lang.IllegalArgumentException
at com.google.common.base.Preconditions.checkArgument(Preconditions.java:108)

at org.apache.beam.sdk.io.gcp.bigquery.BatchLoads.expandTriggered(BatchLoads.java:209)

at org.apache.beam.sdk.io.gcp.bigquery.BatchLoads.expand(BatchLoads.java:546)
at org.apache.beam.sdk.io.gcp.bigquery.BatchLoads.expand(BatchLoads.java:79)

at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:537)
at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:472)

at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:286)
at org.apache.beam.sdk.io.gcp.bigquery.BigQueryIO$Write.expandTyped(BigQueryIO.java:1550)

at org.apache.beam.sdk.io.gcp.bigquery.BigQueryIO$Write.expand(BigQueryIO.java:1497)

at org.apache.beam.sdk.io.gcp.bigquery.BigQueryIO$Write.expand(BigQueryIO.java:980)
at org.apache.beam.sdk.Pipeline.applyInternal(Pipeline.java:537)

at org.apache.beam.sdk.Pipeline.applyTransform(Pipeline.java:491)
at org.apache.beam.sdk.values.PCollection.apply(PCollection.java:299)

at com.travelaudience.data.job.rtbtobigquery.Main$.main(Main.scala:74)
at com.travelaudience.data.job.rtbtobigquery.Main.main(Main.scala)

```

Either default _numFileShards_ should be used or it should be precised in the documentation that this has to be set.

Imported from Jira [BEAM-3766](https://issues.apache.org/jira/browse/BEAM-3766). Original Jira may contain additional context.
Reported by: benjben.

Start with the FILE_LOADS documentation for BigQueryIO.Write.Method and the BatchLoads.expandTriggered check shown in the stack trace; compare it with BigQueryIO.java's write configuration. Done means the documentation clearly states whether withNumFileShards is required for unbounded FILE_LOADS inputs or a default is supported.

Written by the indexing model from the issue text.