v0.2.13

Explicitly use local FS when writing parquet files: #40

v0.2.12

stale watermark after partition revoked fixed

#36

v0.2.11

#35

0.2.10

remove metadata size histogram - #34

0.2.9

Use full record range in partitioned record batch: #33

0.2.8 - Add batched record metric

Add batched record metric - #31

v0.2.7

Fix recovery bug in HDFS loader - #29

v0.2.6

#28
Reduce Kafka offset commit metadata metric size

V0.2.5

#23

• Kafka Metadata compression

#22

• Bumped build to JDK 11
• Bump all GitHub action plugins
• Upgraded sbt, fixed all warnings in build
• Switched to coursier
• Upgraded s3 sdk
• Switched to official Vertica docker image / jdbc driver, bumped them
• Switched to bitnami/kafka docker image, removed zookeeper, moved to KRaft mode
• Upgraded clickhouse-jdbc and switched to com.clickhouse package
• Added support for compression in ClickHouse
• Switched from TrieMap to mutable.HashMap in all single-threaded code
• Reverted to the watermark.delay.ms metric instead of watermark.ms
• Made internal Metrics objects private

Download: https://search.maven.org/artifact/com.adform/stream-loader-hadoop_2.13/0.2.5/jar

v0.2.4

Fix partition assignment bug and refactor.