Fix incorrect computation of reader columns for S3SelectRecordCursor

[phd3]

cc @findepi

A couple of tests in hive-tests have been failing with the following stacktrace, since #1720 was checked in.

java.lang.ArrayIndexOutOfBoundsException: 0
	at io.prestosql.plugin.hive.GenericHiveRecordCursor.isNull(GenericHiveRecordCursor.java:464)
	at io.prestosql.plugin.hive.HiveCoercionRecordCursor.isNull(HiveCoercionRecordCursor.java:155)
	at io.prestosql.plugin.hive.HiveRecordCursor.isNull(HiveRecordCursor.java:239)
	at io.prestosql.spi.connector.RecordPageSource.getNextPage(RecordPageSource.java:98)
	at io.prestosql.testing.MaterializedResult.materializeSourceDataStream(MaterializedResult.java:420)
	at io.prestosql.plugin.hive.AbstractTestHiveFileSystem.createTable(AbstractTestHiveFileSystem.java:454)
	at io.prestosql.plugin.hive.AbstractTestHiveFileSystem.testTableCreation(AbstractTestHiveFileSystem.java:389)
	at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
	at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
	at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
	at java.lang.reflect.Method.invoke(Method.java:498)
	at org.testng.internal.MethodInvocationHelper.invokeMethod(MethodInvocationHelper.java:104)
	at org.testng.internal.Invoker.invokeMethod(Invoker.java:645)
	at org.testng.internal.Invoker.invokeTestMethod(Invoker.java:851)
	at org.testng.internal.Invoker.invokeTestMethods(Invoker.java:1177)
	at org.testng.internal.TestMethodWorker.invokeTestMethods(TestMethodWorker.java:129)
	at org.testng.internal.TestMethodWorker.run(TestMethodWorker.java:112)
	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
	at java.lang.Thread.run(Thread.java:748)

(These tests are not executed for PRs, since tests in hive-tests require usage of AWS_ACCESS_KEY_ID defined in secrets. Right now it looks like the usage of secrets is not allowed for PRs coming from a forked repo. So tests on the PR succeed, but the commits in master are failing these tests.)

The stacktrace points to GenericHiveRecordCursor failing to get the first value in loaded variable. This indicates that GenericHiveRecordCursor was initialized with 0 columns. (Ref) This seems to happen because of a bug in S3SelectRecordCursorProvider. In case when there is no explicit projection required, an empty list of columns was being used, rather than passing along the input columns.

[findepi]

I posted #3450

[martint]

Tests passed in #3450