pabloem commented on a change in pull request #12151: URL: https://github.com/apache/beam/pull/12151#discussion_r454524365
########## File path: sdks/java/io/snowflake/src/main/java/org/apache/beam/sdk/io/snowflake/SnowflakeIO.java ########## @@ -684,14 +819,61 @@ private void checkArguments() { (getDataSourceProviderFn() != null), "withDataSourceConfiguration() or withDataSourceProviderFn() is required"); - checkArgument(getTable() != null, "withTable() is required"); + if (input.isBounded() == PCollection.IsBounded.UNBOUNDED) { + checkArgument(getSnowPipe() != null, "withSnowPipe() is required"); + } else { + checkArgument(getTable() != null, "to() is required"); + } } - private PCollection<String> write(PCollection<T> input, String stagingBucketDir) { + private PCollection<T> writeStream(PCollection<T> input, String stagingBucketDir) { SnowflakeService snowflakeService = - getSnowflakeService() != null ? getSnowflakeService() : new SnowflakeServiceImpl(); + getSnowflakeService() != null + ? getSnowflakeService() + : new SnowflakeStreamingServiceImpl(); + + /* Ensure that files will be created after specific record count or duration specified */ + PCollection<T> inputInGlobalWindow = + input.apply( + "rewindowIntoGlobal", + Window.<T>into(new GlobalWindows()) + .triggering( + Repeatedly.forever( + AfterFirst.of( + AfterProcessingTime.pastFirstElementInPane() + .plusDelayOf(getFlushTimeLimit()), + AfterPane.elementCountAtLeast(getFlushRowLimit())))) + .discardingFiredPanes()); + + int shards = (getShardsNumber() > 0) ? getShardsNumber() : DEFAULT_STREAMING_SHARDS_NUMBER; + PCollection files = writeFiles(inputInGlobalWindow, stagingBucketDir, shards); + + /* Ensuring that files will be ingested after flush time */ + files = + (PCollection) + files.apply( + "applyUserTrigger", + Window.<T>into(new GlobalWindows()) + .triggering( + Repeatedly.forever( + AfterProcessingTime.pastFirstElementInPane() + .plusDelayOf(getFlushTimeLimit()))) + .discardingFiredPanes()); + files = + (PCollection) + files.apply( + "Create list of logs to copy", + Combine.globally(new Concatenate()).withoutDefaults()); Review comment: You can just KeyBy(same key) + GroupByKey, and read one by one - but it may be fine to leave as is if we don't expect > 100 files or so ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org