Provide better error message in ReadSessionCreator.create() if table doesn't exist

[jphalip]

You you call the ReadSessionCreator.create() method and the table doesn't exist, you get a NullPointerException. This is because the ReadSessionCreator class in the bigquery-connector-common library does not set the setThrowNotFound option when it calls the BQ API's getTable() method. We should maybe modify that code to better handle this case and provide a better error message instead of just throwing a NullPointerException.

[davidrabinowitz]

That's because we check it earlier, this is the error in spark:

val df = spark.read.format("bigquery").load("davidrab.doesnotexist")
java.lang.RuntimeException: Table project.dataset.doesnotexist not found
  at scala.sys.package$.error(package.scala:30)
  at com.google.cloud.spark.bigquery.BigQueryRelationProvider.$anonfun$createRelationInternal$1(BigQueryRelationProvider.scala:80)
  at scala.Option.getOrElse(Option.scala:189)
  at com.google.cloud.spark.bigquery.BigQueryRelationProvider.createRelationInternal(BigQueryRelationProvider.scala:80)
  at com.google.cloud.spark.bigquery.BigQueryRelationProvider.createRelation(BigQueryRelationProvider.scala:46)
  at org.apache.spark.sql.execution.datasources.DataSource.resolveRelation(DataSource.scala:355)
  at org.apache.spark.sql.DataFrameReader.loadV1Source(DataFrameReader.scala:325)
  at org.apache.spark.sql.DataFrameReader.$anonfun$load$3(DataFrameReader.scala:307)
  at scala.Option.getOrElse(Option.scala:189)
  at org.apache.spark.sql.DataFrameReader.load(DataFrameReader.scala:307)
  at org.apache.spark.sql.DataFrameReader.load(DataFrameReader.scala:239)
  ... 47 elided

[jphalip]

Would it make sense to still set the setThrowNotFound option? That shouldn't break the Spark code (since it already checks the table's existence prior), and would allow the Hive connector to get a better error message without having to make a separate API call to check the table's existence.

[jphalip]

I ended up making an extra call to BigQuery to check the table's existence upfront: https://github.com/GoogleCloudDataproc/hive-bigquery-connector/commit/dafa09f2177b4134c124a81908a5171c12987580

Would still be nice to update ReadSessionCreator.create() to directly handle that edge case so we can save making that extra call.