You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "jingz-db (via GitHub)" <gi...@apache.org> on 2024/03/25 21:43:08 UTC
[PR] Integrate range scan encoder changes with timer implementation [spark]
jingz-db opened a new pull request, #45709:
URL: https://github.com/apache/spark/pull/45709
<!--
Thanks for sending a pull request! Here are some tips for you:
1. If this is your first time, please read our contributor guidelines: https://spark.apache.org/contributing.html
2. Ensure you have added or run the appropriate tests for your PR: https://spark.apache.org/developer-tools.html
3. If the PR is unfinished, add '[WIP]' in your PR title, e.g., '[WIP][SPARK-XXXX] Your PR title ...'.
4. Be sure to keep the PR description updated to reflect all changes.
5. Please write your PR title to summarize what this PR proposes.
6. If possible, provide a concise example to reproduce the issue for a faster review.
7. If you want to add a new configuration, please read the guideline first for naming configurations in
'core/src/main/scala/org/apache/spark/internal/config/ConfigEntry.scala'.
8. If you want to add or modify an error type or message, please read the guideline first in
'common/utils/src/main/resources/error/README.md'.
-->
### What changes were proposed in this pull request?
<!--
Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue.
If possible, please consider writing useful notes for better and faster reviews in your PR. See the examples below.
1. If you refactor some codes with changing classes, showing the class hierarchy will help reviewers.
2. If you fix some SQL features, you can provide some references of other DBMSes.
3. If there is design documentation, please add the link.
4. If there is a discussion in the mailing list, please add the link.
-->
Previously timer state implementation was using No prefix rocksdb state encoder. When doing `iterator()` or `prefix()`, the returned iterator is not sorted on timestamp value. After Anish's PR for supporting range scan encoder, we could integrate it with `TimerStateImpl` such that we will use range scan encoder on `timer to key`.
### Why are the changes needed?
<!--
Please clarify why the changes are needed. For instance,
1. If you propose a new API, clarify the use case for a new API.
2. If you fix a bug, you can clarify why it is a bug.
-->
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as the documentation fix.
If yes, please clarify the previous behavior and the change this PR proposes - provide the console output, description and/or an example to show the behavior difference if possible.
If possible, please also clarify if this is a user-facing change compared to the released Spark versions or within the unreleased branches such as master.
If no, write 'No'.
-->
No
### How was this patch tested?
<!--
If tests were added, say they were added here. Please make sure to add some test cases that check the changes thoroughly including negative and positive cases if possible.
If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future.
If tests were not added, please describe why they were not added and/or why it was difficult to add.
If benchmark tests were added, please run the benchmarks in GitHub Actions for the consistent environment, and the instructions could accord to: https://spark.apache.org/developer-tools.html#github-workflow-benchmarks.
-->
Added unit tests in `TimerSuite`
### Was this patch authored or co-authored using generative AI tooling?
<!--
If generative AI tooling has been used in the process of authoring this patch, please include the
phrase: 'Generated-by: ' followed by the name of the tool and its version.
If no, write 'No'.
Please refer to the [ASF Generative Tooling Guidance](https://www.apache.org/legal/generative-tooling.html) for details.
-->
No
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "jingz-db (via GitHub)" <gi...@apache.org>.
jingz-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1541998642
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TimerStateImpl.scala:
##########
@@ -188,9 +187,12 @@ class TimerStateImpl(
/**
* Function to get all the registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
Review Comment:
Thanks for pointing this out (I did not realize this before but looks like we did the right thing)! We've also left out those implementation details in the user facing function `getExpiredTimers()` in `StatefulProcessorHandleimpl` here:
https://github.com/jingz-db/spark/blob/36ab93519ae1ab43e20926ae47406cff668a3b10/sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/StatefulProcessorHandleImpl.scala#L166-L174
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on PR #45709:
URL: https://github.com/apache/spark/pull/45709#issuecomment-2021915988
Thanks! Merging to master.
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540180091
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TimerStateImpl.scala:
##########
@@ -188,9 +187,12 @@ class TimerStateImpl(
/**
* Function to get all the registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
Review Comment:
Nit: maybe add a small comment here mentioning that we perform a range scan and stop iterating once the key row timestamp exceeds the threshold
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1539925533
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TransformWithStateExec.scala:
##########
@@ -161,25 +161,23 @@ case class TransformWithStateExec(
assert(batchTimestampMs.isDefined)
val batchTimestamp = batchTimestampMs.get
val procTimeIter = processorHandle.getExpiredTimers()
Review Comment:
Could we pass the timeout to the iterator and just return the expired rows instead ? i.e. move the filtering logic within the `TimerStateImpl` ?
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540147143
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TimerStateImpl.scala:
##########
@@ -188,17 +187,20 @@ class TimerStateImpl(
/**
* Function to get all the registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
+ * will return every timers that has (strictly) smaller timestamp
* @return - iterator of all the registered timers for all grouping keys
*/
- def getExpiredTimers(): Iterator[(Any, Long)] = {
+ def getExpiredTimers(expiryTimestampMs: Long): Iterator[(Any, Long)] = {
+ // this iter is increasingly sorted on timestamp
val iter = store.iterator(tsToKeyCFName)
new NextIterator[(Any, Long)] {
override protected def getNext(): (Any, Long) = {
- if (iter.hasNext) {
- val rowPair = iter.next()
- val keyRow = rowPair.key
- val result = getTimerRowFromSecIndex(keyRow)
+ val rowPair = if (iter.hasNext) iter.next() else null
Review Comment:
Lets keep the original if condition as it is ? and add a condition to return `null` or `result` within this case ?
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540148899
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/StatefulProcessorHandleImpl.scala:
##########
@@ -164,11 +164,13 @@ class StatefulProcessorHandleImpl(
/**
* Function to retrieve all registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
+ * will return every timers that has (strictly) smaller timestamp
Review Comment:
Nit: `will return all timers that have timestamp less than passed threshold`
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540446675
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TimerStateImpl.scala:
##########
@@ -188,9 +187,12 @@ class TimerStateImpl(
/**
* Function to get all the registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
Review Comment:
This class is not user-facing, right? If it is, I'd suggest avoiding implementation detail. Looks like as it doesn't seem to be an user facing, but just to double check.
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540446675
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TimerStateImpl.scala:
##########
@@ -188,9 +187,12 @@ class TimerStateImpl(
/**
* Function to get all the registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
Review Comment:
This class is not user-facing, right? If it is, I'd suggest avoiding implementation detail. Looks like as it doesn't seem to be an user facing, but just to remind.
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540146283
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TransformWithStateExec.scala:
##########
@@ -160,26 +160,18 @@ case class TransformWithStateExec(
case ProcessingTime =>
assert(batchTimestampMs.isDefined)
val batchTimestamp = batchTimestampMs.get
- val procTimeIter = processorHandle.getExpiredTimers()
- procTimeIter.flatMap { case (keyObj, expiryTimestampMs) =>
- if (expiryTimestampMs < batchTimestamp) {
+ processorHandle.getExpiredTimers()
Review Comment:
We need to pass the expiry threshold right ?
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TransformWithStateExec.scala:
##########
@@ -160,26 +160,18 @@ case class TransformWithStateExec(
case ProcessingTime =>
assert(batchTimestampMs.isDefined)
val batchTimestamp = batchTimestampMs.get
- val procTimeIter = processorHandle.getExpiredTimers()
- procTimeIter.flatMap { case (keyObj, expiryTimestampMs) =>
- if (expiryTimestampMs < batchTimestamp) {
+ processorHandle.getExpiredTimers()
+ .flatMap { case (keyObj, expiryTimestampMs) =>
handleTimerRows(keyObj, expiryTimestampMs, processorHandle)
- } else {
- Iterator.empty
}
- }
case EventTime =>
assert(eventTimeWatermarkForEviction.isDefined)
val watermark = eventTimeWatermarkForEviction.get
- val eventTimeIter = processorHandle.getExpiredTimers()
- eventTimeIter.flatMap { case (keyObj, expiryTimestampMs) =>
- if (expiryTimestampMs < watermark) {
+ processorHandle.getExpiredTimers()
Review Comment:
Same here ?
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "HeartSaVioR (via GitHub)" <gi...@apache.org>.
HeartSaVioR closed pull request #45709: [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation
URL: https://github.com/apache/spark/pull/45709
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] Integrate range scan encoder changes with timer implementation [spark]
Posted by "HyukjinKwon (via GitHub)" <gi...@apache.org>.
HyukjinKwon commented on PR #45709:
URL: https://github.com/apache/spark/pull/45709#issuecomment-2019244232
@jingz-db mind liking the JIRA into the PR title? See also https://spark.apache.org/contributing.html
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540147961
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TimerStateImpl.scala:
##########
@@ -188,17 +187,20 @@ class TimerStateImpl(
/**
* Function to get all the registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
+ * will return every timers that has (strictly) smaller timestamp
* @return - iterator of all the registered timers for all grouping keys
*/
- def getExpiredTimers(): Iterator[(Any, Long)] = {
+ def getExpiredTimers(expiryTimestampMs: Long): Iterator[(Any, Long)] = {
+ // this iter is increasingly sorted on timestamp
val iter = store.iterator(tsToKeyCFName)
new NextIterator[(Any, Long)] {
override protected def getNext(): (Any, Long) = {
- if (iter.hasNext) {
- val rowPair = iter.next()
- val keyRow = rowPair.key
- val result = getTimerRowFromSecIndex(keyRow)
+ val rowPair = if (iter.hasNext) iter.next() else null
+ val result: (Any, Long) =
+ if (rowPair != null) getTimerRowFromSecIndex(rowPair.key) else null
+ if (result != null && result._2 < expiryTimestampMs) {
Review Comment:
Lets do `<=` comparison so that the timer doesn't have to wait for the next batch to fire ?
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540170826
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/streaming/state/TimerSuite.scala:
##########
@@ -99,15 +102,69 @@ class TimerSuite extends StateVariableSuiteBase {
ImplicitGroupingKeyTracker.removeImplicitKey()
ImplicitGroupingKeyTracker.setImplicitKey("test_key1")
- assert(timerState1.getExpiredTimers().toSet ===
- Set(("test_key2", 15000L), ("test_key1", 2000L), ("test_key1", 1000L)))
+ assert(timerState1.getExpiredTimers(Long.MaxValue).toSeq ===
+ Seq(("test_key1", 1000L), ("test_key1", 2000L), ("test_key2", 15000L)))
+ assert(timerState1.getExpiredTimers(10000L).toSeq ===
+ Seq(("test_key1", 1000L), ("test_key1", 2000L)))
assert(timerState1.listTimers().toSet === Set(1000L, 2000L))
ImplicitGroupingKeyTracker.removeImplicitKey()
ImplicitGroupingKeyTracker.setImplicitKey("test_key2")
assert(timerState2.listTimers().toSet === Set(15000L))
- assert(timerState2.getExpiredTimers().toSet ===
- Set(("test_key2", 15000L), ("test_key1", 2000L), ("test_key1", 1000L)))
+ assert(timerState2.getExpiredTimers(1500L).toSeq === Seq(("test_key1", 1000L)))
+ }
+ }
+
+ testWithTimeOutMode("Range scan on second index timer key - " +
+ "verify timestamp is sorted for single instance") { timeoutMode =>
+ tryWithProviderResource(newStoreProviderWithStateVariable(true)) { provider =>
+ val store = provider.getStore(0)
+
+ ImplicitGroupingKeyTracker.setImplicitKey("test_key")
+ val timerState = new TimerStateImpl(store, timeoutMode,
+ Encoders.STRING.asInstanceOf[ExpressionEncoder[Any]])
+ val timerTimerstamps = Seq(931L, 8000L, 452300L, 4200L, 90L, 1L, 2L, 8L, 3L, 35L, 6L, 9L, 5L)
+ // register/put unordered timestamp into rocksDB
+ timerTimerstamps.foreach(timerState.registerTimer)
+ assert(timerState.getExpiredTimers(Long.MaxValue).toSeq.map(_._2) === timerTimerstamps.sorted)
+ assert(timerState.getExpiredTimers(4200L).toSeq.map(_._2) ===
+ timerTimerstamps.sorted.takeWhile(_ < 4200L))
+ assert(timerState.getExpiredTimers(Long.MinValue).toSeq === Seq.empty)
+ ImplicitGroupingKeyTracker.removeImplicitKey()
+ }
+ }
+
+ testWithTimeOutMode("test range scan on second index timer key - " +
+ "verify timestamp is sorted for multiple instances") { timeoutMode =>
+ tryWithProviderResource(newStoreProviderWithStateVariable(true)) { provider =>
+ val store = provider.getStore(0)
+
+ ImplicitGroupingKeyTracker.setImplicitKey("test_key1")
+ val timerState1 = new TimerStateImpl(store, timeoutMode,
+ Encoders.STRING.asInstanceOf[ExpressionEncoder[Any]])
+ val timerTimestamps1 = Seq(64L, 32L, 1024L, 4096L, 0L, 1L)
+ timerTimestamps1.foreach(timerState1.registerTimer)
+
+ val timerState2 = new TimerStateImpl(store, timeoutMode,
+ Encoders.STRING.asInstanceOf[ExpressionEncoder[Any]])
+ val timerTimestamps2 = Seq(931L, 8000L, 452300L, 4200L)
+ timerTimestamps2.foreach(timerState2.registerTimer)
+ ImplicitGroupingKeyTracker.removeImplicitKey()
+
+ ImplicitGroupingKeyTracker.setImplicitKey("test_key3")
+ val timerState3 = new TimerStateImpl(store, timeoutMode,
+ Encoders.STRING.asInstanceOf[ExpressionEncoder[Any]])
+ val timerTimerStamps3 = Seq(1L, 2L, 8L, 3L)
+ timerTimerStamps3.foreach(timerState3.registerTimer)
+ ImplicitGroupingKeyTracker.removeImplicitKey()
+
+ ImplicitGroupingKeyTracker.setImplicitKey("test_key1")
Review Comment:
If we remove setting the implicit key just for this section (for the calls to getExpiredTimers) would the test still work ?
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "jingz-db (via GitHub)" <gi...@apache.org>.
jingz-db commented on PR #45709:
URL: https://github.com/apache/spark/pull/45709#issuecomment-2021101002
> @jingz-db mind liking the JIRA into the PR title? See also https://spark.apache.org/contributing.html
Just did. Thanks!
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "anishshri-db (via GitHub)" <gi...@apache.org>.
anishshri-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540147961
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TimerStateImpl.scala:
##########
@@ -188,17 +187,20 @@ class TimerStateImpl(
/**
* Function to get all the registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
+ * will return every timers that has (strictly) smaller timestamp
* @return - iterator of all the registered timers for all grouping keys
*/
- def getExpiredTimers(): Iterator[(Any, Long)] = {
+ def getExpiredTimers(expiryTimestampMs: Long): Iterator[(Any, Long)] = {
+ // this iter is increasingly sorted on timestamp
val iter = store.iterator(tsToKeyCFName)
new NextIterator[(Any, Long)] {
override protected def getNext(): (Any, Long) = {
- if (iter.hasNext) {
- val rowPair = iter.next()
- val keyRow = rowPair.key
- val result = getTimerRowFromSecIndex(keyRow)
+ val rowPair = if (iter.hasNext) iter.next() else null
+ val result: (Any, Long) =
+ if (rowPair != null) getTimerRowFromSecIndex(rowPair.key) else null
+ if (result != null && result._2 < expiryTimestampMs) {
Review Comment:
Lets do `<=` comparison so that the timer doesn't have to wait for the next batch to fire ?
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "jingz-db (via GitHub)" <gi...@apache.org>.
jingz-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1540174052
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/streaming/state/TimerSuite.scala:
##########
@@ -99,15 +102,69 @@ class TimerSuite extends StateVariableSuiteBase {
ImplicitGroupingKeyTracker.removeImplicitKey()
ImplicitGroupingKeyTracker.setImplicitKey("test_key1")
- assert(timerState1.getExpiredTimers().toSet ===
- Set(("test_key2", 15000L), ("test_key1", 2000L), ("test_key1", 1000L)))
+ assert(timerState1.getExpiredTimers(Long.MaxValue).toSeq ===
+ Seq(("test_key1", 1000L), ("test_key1", 2000L), ("test_key2", 15000L)))
+ assert(timerState1.getExpiredTimers(10000L).toSeq ===
+ Seq(("test_key1", 1000L), ("test_key1", 2000L)))
assert(timerState1.listTimers().toSet === Set(1000L, 2000L))
ImplicitGroupingKeyTracker.removeImplicitKey()
ImplicitGroupingKeyTracker.setImplicitKey("test_key2")
assert(timerState2.listTimers().toSet === Set(15000L))
- assert(timerState2.getExpiredTimers().toSet ===
- Set(("test_key2", 15000L), ("test_key1", 2000L), ("test_key1", 1000L)))
+ assert(timerState2.getExpiredTimers(1500L).toSeq === Seq(("test_key1", 1000L)))
+ }
+ }
+
+ testWithTimeOutMode("Range scan on second index timer key - " +
+ "verify timestamp is sorted for single instance") { timeoutMode =>
+ tryWithProviderResource(newStoreProviderWithStateVariable(true)) { provider =>
+ val store = provider.getStore(0)
+
+ ImplicitGroupingKeyTracker.setImplicitKey("test_key")
+ val timerState = new TimerStateImpl(store, timeoutMode,
+ Encoders.STRING.asInstanceOf[ExpressionEncoder[Any]])
+ val timerTimerstamps = Seq(931L, 8000L, 452300L, 4200L, 90L, 1L, 2L, 8L, 3L, 35L, 6L, 9L, 5L)
+ // register/put unordered timestamp into rocksDB
+ timerTimerstamps.foreach(timerState.registerTimer)
+ assert(timerState.getExpiredTimers(Long.MaxValue).toSeq.map(_._2) === timerTimerstamps.sorted)
+ assert(timerState.getExpiredTimers(4200L).toSeq.map(_._2) ===
+ timerTimerstamps.sorted.takeWhile(_ < 4200L))
+ assert(timerState.getExpiredTimers(Long.MinValue).toSeq === Seq.empty)
+ ImplicitGroupingKeyTracker.removeImplicitKey()
+ }
+ }
+
+ testWithTimeOutMode("test range scan on second index timer key - " +
+ "verify timestamp is sorted for multiple instances") { timeoutMode =>
+ tryWithProviderResource(newStoreProviderWithStateVariable(true)) { provider =>
+ val store = provider.getStore(0)
+
+ ImplicitGroupingKeyTracker.setImplicitKey("test_key1")
+ val timerState1 = new TimerStateImpl(store, timeoutMode,
+ Encoders.STRING.asInstanceOf[ExpressionEncoder[Any]])
+ val timerTimestamps1 = Seq(64L, 32L, 1024L, 4096L, 0L, 1L)
+ timerTimestamps1.foreach(timerState1.registerTimer)
+
+ val timerState2 = new TimerStateImpl(store, timeoutMode,
+ Encoders.STRING.asInstanceOf[ExpressionEncoder[Any]])
+ val timerTimestamps2 = Seq(931L, 8000L, 452300L, 4200L)
+ timerTimestamps2.foreach(timerState2.registerTimer)
+ ImplicitGroupingKeyTracker.removeImplicitKey()
+
+ ImplicitGroupingKeyTracker.setImplicitKey("test_key3")
+ val timerState3 = new TimerStateImpl(store, timeoutMode,
+ Encoders.STRING.asInstanceOf[ExpressionEncoder[Any]])
+ val timerTimerStamps3 = Seq(1L, 2L, 8L, 3L)
+ timerTimerStamps3.foreach(timerState3.registerTimer)
+ ImplicitGroupingKeyTracker.removeImplicitKey()
+
+ ImplicitGroupingKeyTracker.setImplicitKey("test_key1")
Review Comment:
Yes it works. Will removed them in the next commit.
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-47570][SS] Integrate range scan encoder changes with timer implementation [spark]
Posted by "jingz-db (via GitHub)" <gi...@apache.org>.
jingz-db commented on code in PR #45709:
URL: https://github.com/apache/spark/pull/45709#discussion_r1541998642
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/TimerStateImpl.scala:
##########
@@ -188,9 +187,12 @@ class TimerStateImpl(
/**
* Function to get all the registered timers for all grouping keys
+ * @param expiryTimestampMs Threshold for expired timestamp in milliseconds, this function
Review Comment:
Thanks for pointing this out (I did not realize this before but looks like we did the right thing)!
--
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.
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org