pabloem commented on a change in pull request #12151:
URL: https://github.com/apache/beam/pull/12151#discussion_r451294464
##########
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());
Review comment:
A PCollection after a combine / GBK / etc - keeps the trigger
configuration from upstream. Did you find that you needed to restate the
trigger like this?
----------------------------------------------------------------
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:
[email protected]
