[TEMP] [WIP] Transaction manager support

[ajantha-bhat]

Why is this PR needed?

Based on the discussion and design document from https://issues.apache.org/jira/browse/CARBONDATA-4171

What changes were proposed in this PR?

• Added a new ACID module.
• Defined a transaction manager interface (can support db and table) and implemented a file-based transaction manager to write the transaction log file.
• Defined a segment store interface (can support db and table) and implemented a file-based segment store to write the table status with the transaction id (multiple table status file)
• Handled the basic transaction conflict scenario for the file-based segment store.
• Lock support for transaction log file

TODO:

1. Write wrappers for the segment store interface (segment status updater and reader) and replace them in the current code.
2. Remove table status lock everywhere and changing segment id logic
3. Adding segment id as the return value of each transaction operation to fill the affected segments
4. similar to segment store interface (table status), need to handle for update status file and segment file also
5. Try to decouple data and metadata operation for each transaction command (may be write new classes), so metadata can be supported for other open formats also.
6. Test optimistic concurrency by running parallel transactions (update, compation) and cover all the scenarios.
   On top of these, we can provide time travel SQL support and multi-format support. Across table transaction support

Does this PR introduce any user interface change?

• No

Is any new test case added?

• No
• Yes

[CarbonDataQA2]

Build Failed with Spark 2.4.5, Please check CI http://121.244.95.60:12602/job/ApacheCarbon_PR_Builder_2.4.5/3706/

[CarbonDataQA2]

Build Failed with Spark 2.3.4, Please check CI http://121.244.95.60:12602/job/ApacheCarbonPRBuilder2.3/5450/