[WIP][SPARK-20824][SQL] Generate code to get value from table cache with wider column in ColumnarBatch

[kiszk]

What changes were proposed in this pull request?

This PR waits for merging #18066

This PR adds compensation code for disabling the whole-stage codegen due to large number of columns.

When the whole-stage codegen is disabled, an iterator gets data from a row. Since the data is stored in columnar sotrage CachedBatch, the iterator is generated to copy data from the columnar storage to a row.
This PR supports only interger and double types. Another PR will support other data types.

How was this patch tested?

Added test suites into InMemoryColumnarQuerySuite

[SparkQA]

Test build #77715 has finished for PR 18194 at commit 0c22730.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
• class SpecificColumnarIterator extends $

[kiszk]

Jenkins, retest this please

[SparkQA]

Test build #77716 has finished for PR 18194 at commit 0c22730.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
• class SpecificColumnarIterator extends $

[gatorsmile]

@kiszk Could we close this PR? and you can re-submit it later.

[kiszk]

Sure, it will not be used.