Release v1.0.21

• Fix Parquet Extract case issue on column types
• Fix SQLite JDBCExtract not using limit...offset syntax

Release v1.0.20

• Add support for reading from/writing to a directory containing Parquet Files (e.g. on HDFS)
• Add support for defining JDBC idle timeouts (and auto-refresh intervals) for connection pools, for better stability when doing larger batch sizes.
  • Add optional idleTimeout configuration parameter to both JDBCExtract and JDBCLoad, for a timeout in seconds. Default if not supplied is 0, meaning no idle timeout.

Release v1.0.19

• Better File Output Options for CSV and JSONL (newline-delimited JSON objects)
  • Usage is not recommended in production environments
  • Please refer to the example configs for CSV and JSONL

Thanks to @nishanthpp93 for his contributions to this release

Release v1.0.18

• Adds support for HCatalog Data Sources, e.g. Apache Hive (Extract Only)
  • Expected Configuration: {"metastore_uris": "thrift://metastore-host:port", "database": "source_database_name", "table": "source_table_name"}

Release v1.0.17

• Downgrade packaged flink version to 1.13 for dependency version match

Release v1.0.16

Fix identifier_col param load order to before the orderby cols are determined

Release v1.0.15

Fix identifier_col configuration option not being loaded

Release v1.0.14

• Update backing beam version to latest (2.36.0) to support updated featureset

Release v1.0.13

Add Solr Extract Component

Expected configuration can be found at https://raw.githubusercontent.com/OHNLP/Backbone/master/IO/src/main/java/org/ohnlp/backbone/io/solr/SolrExtract.java

and further modification can be done based on user needs.

Release v1.0.12

Interim release deprecated by v1.0.13