Server-to-Parallel - S2PX Release v2.0

MCI-6471

Improved command usage notes

MCI-6466

Added TeradataConnector and associated stages covered by CCMT to the supported list

MCI-6180

Fix Hash job count reference

MCI-5217

Add new query to detect Link Partition using Hash Partitioning Algorithm

MCI-6509

HashedFile columns are now unbounded during read

MCI-6763

• Move checking Aggregator column isGroup to a method

• Fix translation of Server to PX Aggregator Group Key (OEM Skill Case TS020256049)

MCI-6497

Added tests for JobReferenceTypeAnalyser (processHashedFile)

MCI-6703

• Fix bug in shared container parsing logic

• Add support for hashed file accounts

MCI-6724

• Refactor DataSetHashedFileMigrator to make is easier to replace dataset stage with another implementation

• Add working implementation for generating Hashed File -> DataSet migration jobs

• Fix incorrectly identifiers for fully generated shared containers, add support for server shared containers in px jobs/containers

MCI-6705

Disabled Hashed File Migration and Hashed File parameters when using DataSets instead of DRS Connectors

MCI-6701

• Handle write mode and fail when columns are missing

• Add simple feature toggle for replacing hashed files with dataset stages

MCI-6703

Split InsertedHashedFileLookupConversion into distinct DRS and DataSet implementations.

MCI-6497

Add Hashed File Lookup mode to config

MCI-6498

• Fix intermittent fork-join breaker test failure. Caused by scores not being correctly adjusted while breaking fork-joins.

• Refactor existing feature toggle and add workaround for ODBC connector crash

• Resolve DecimalToString typecast differences and preserve trailing zeros when reading decimals from connectors

• Change link name to avoid name collection which breaks S2PX.

• Add DataStage Unit Test demonstrating Decimal to String problems

MCI-6471

Updates usage notes

MCI-6465

Added TeradataConnector translation and table action

MCI-6464

Removed explicit DecimalToDFloat() conversion for function and routine arguments.

MCI-6458

Configure connector based type inferrencing for both Teradata and Netezza connectors.

MCI-6250

• Made sure UT converter uses application level precision and scale

• Corrected datastage test case for aggregate by date which seems to have been modified to pass in px when behaviour would be different in server

• Addressed not handling implicit type conversions that dont actually require type casts for output links

MCI-6350

Corrected minor mistake on expected output

MCI-6301

Fixed bug when decomposing job with no split groups

MCI-6241

Added provided dependency on debug-command-shell

MCI-6240

• Only store name of failed containers

• Set non-zero return code when exception converting unit test

MCI-6271

Removed unit tests that stub stages in SCs

MCI-6236

Fixed NullPointerException

MCI-6186

• Change Fork-Join extraction algorithm from identifying shortest fork-join cycles of all varieties to exhaustively identifying all partially broken fork-joins and then all unbroken fork-joins

• Update fork-join breaker algorithm to handle multiple loop permutations and fix bugs such as the identification and handling of ‘unbreakable’ OutputPins

• Update ForkJoin data structures to allow equality checks (to support duplicate removal) and human readable string representation

• Fix decomposition optimization which was failing to iteratively merge decomposition when minimizing jobs

MCI-6070

• Add unit test to verify that the complex job provided by a customer can in fact be decomposed.

• Updated Fork-Join search algorithm to exclude loops where forks or joins can be split, prevents short superfluos loops from longer critical loops form being detected

• Prevent broken decomposition when Fork stages have multiple output links to the same next stage.

<TBC>

Fix Hashed File DRS Connectors using ORACLE instead of Oracle as DBMS type.

MCI-6097

Duplicate group and calc columns prior to iteration to prevent concurrent modification exceptions