I'm trying to run a pySpark job with custom inputs, for testing purposes. The job has three sets of input, each read from a table in a different metastore database.
The data is read in spark with: hiveContext.table('myDb.myTable')
The test inputs are three files. In an attempt to not change any of the original code, I read all three inputs into DataFrame
s, and attempt to register a temp table with myDF.registerTempTable('myDb.myTable')
.
The problem is that spark fails with org.apache.spark.sql.catalyst.analysis.NoSuchTableException
.
I've also tried:
hiveContext.sql('create database if not exists myDb')
hiveContext.sql('use myDb')
myDF.registerTempTable('myTable')
But that fails as well.
Any idea why the table cannot be found?
Using Spark 1.6