You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "sadikovi (via GitHub)" <gi...@apache.org> on 2024/03/19 01:28:47 UTC
[PR] [SPARK-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
sadikovi opened a new pull request, #45578:
URL: https://github.com/apache/spark/pull/45578
<!--
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.
-->
This PR fixes a regression introduced by [SPARK-46633](https://issues.apache.org/jira/browse/SPARK-46633), commit: https://github.com/apache/spark/commit/3a6b9adc21c25b01746cc31f3b75fe061a63204c where one could not read an empty Avro file as the reader would be stuck in an infinite loop.
I reverted the code to the pre-SPARK-46633 version and added handling for empty blocks - `hasNext` status could return false but could still trigger loading of the next block so we need to call the method again to get the actual `hasNext` value.
### 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.
-->
Fixes a regression introduced in [SPARK-46633](https://issues.apache.org/jira/browse/SPARK-46633).
### 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.
-->
I added a unit test to verify that empty files can be read correctly.
### 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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "cloud-fan (via GitHub)" <gi...@apache.org>.
cloud-fan commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529914252
##########
connector/avro/src/main/scala/org/apache/spark/sql/avro/AvroUtils.scala:
##########
@@ -197,19 +197,21 @@ private[sql] object AvroUtils extends Logging {
def hasNextRow: Boolean = {
while (!completed && currentRow.isEmpty) {
- if (fileReader.pastSync(stopPosition)) {
+ // In the case of empty blocks in an Avro file, `blockRemaining` could still read as 0 so
+ // `fileReader.hasNext()` returns false but advances the cursor to the next block, so we
+ // need to call `fileReader.hasNext()` again to correctly report if the next record
+ // exists.
Review Comment:
hmmm, what if there are continuous empty blocks? how many times we need to call `hasNext`?
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "cloud-fan (via GitHub)" <gi...@apache.org>.
cloud-fan closed pull request #45578: [SPARK-46990][SQL] Fix loading empty Avro files emitted by event-hubs
URL: https://github.com/apache/spark/pull/45578
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "cloud-fan (via GitHub)" <gi...@apache.org>.
cloud-fan commented on PR #45578:
URL: https://github.com/apache/spark/pull/45578#issuecomment-2009876142
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "sadikovi (via GitHub)" <gi...@apache.org>.
sadikovi commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529737239
##########
connector/avro/src/main/scala/org/apache/spark/sql/avro/AvroUtils.scala:
##########
@@ -197,19 +197,20 @@ private[sql] object AvroUtils extends Logging {
def hasNextRow: Boolean = {
while (!completed && currentRow.isEmpty) {
- if (fileReader.pastSync(stopPosition)) {
+ // In some cases of empty blocks in an Avro file, `fileReader.hasNext()` returns false but
Review Comment:
I have updated the comments to elaborate on this a bit more.
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "sadikovi (via GitHub)" <gi...@apache.org>.
sadikovi commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1531179747
##########
connector/avro/src/main/scala/org/apache/spark/sql/avro/AvroUtils.scala:
##########
@@ -197,19 +197,21 @@ private[sql] object AvroUtils extends Logging {
def hasNextRow: Boolean = {
while (!completed && currentRow.isEmpty) {
- if (fileReader.pastSync(stopPosition)) {
+ // In the case of empty blocks in an Avro file, `blockRemaining` could still read as 0 so
+ // `fileReader.hasNext()` returns false but advances the cursor to the next block, so we
+ // need to call `fileReader.hasNext()` again to correctly report if the next record
+ // exists.
Review Comment:
That's what I was thinking but I could not find an example that double hasNext has not fixed. So it seems to be fine. I just could not find a better way of solving this issue at the moment given the FileReader interface.
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "gengliangwang (via GitHub)" <gi...@apache.org>.
gengliangwang commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529711554
##########
connector/avro/src/test/scala/org/apache/spark/sql/avro/AvroSuite.scala:
##########
@@ -2744,6 +2744,16 @@ abstract class AvroSuite
}
}
}
+
+ test("SPARK-46990: read an empty file where pastSync returns false at EOF") {
Review Comment:
It's ok. Probably my env issue. Mark this as resolved.
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "sadikovi (via GitHub)" <gi...@apache.org>.
sadikovi commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529652672
##########
connector/avro/src/main/scala/org/apache/spark/sql/avro/AvroUtils.scala:
##########
@@ -197,19 +197,19 @@ private[sql] object AvroUtils extends Logging {
def hasNextRow: Boolean = {
while (!completed && currentRow.isEmpty) {
- if (fileReader.pastSync(stopPosition)) {
+ // In some cases of empty blocks in an Avro file, `fileReader.hasNext()` returns false but
+ // advances the cursor to the next block, so we need to call `fileReader.hasNext()` again
+ // to correctly report whether the next record exists.
+ val r = (fileReader.hasNext || fileReader.hasNext) && !fileReader.pastSync(stopPosition)
Review Comment:
Thanks. I updated the code but I used `moreData` variable instead and `if (!moreData)`.
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "sadikovi (via GitHub)" <gi...@apache.org>.
sadikovi commented on PR #45578:
URL: https://github.com/apache/spark/pull/45578#issuecomment-2005800625
Yes, to be honest, I could not find a better solution that would fix empty file and the empty blocks issue. I think this should probably be fixed in Avro. @gengliangwang @cloud-fan How possible is it to fix in Avro and then upgrade the dependency in Spark? 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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "sadikovi (via GitHub)" <gi...@apache.org>.
sadikovi commented on PR #45578:
URL: https://github.com/apache/spark/pull/45578#issuecomment-2005572317
cc @cloud-fan @HyukjinKwon
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "gengliangwang (via GitHub)" <gi...@apache.org>.
gengliangwang commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529687598
##########
connector/avro/src/test/scala/org/apache/spark/sql/avro/AvroSuite.scala:
##########
@@ -2744,6 +2744,16 @@ abstract class AvroSuite
}
}
}
+
+ test("SPARK-46990: read an empty file where pastSync returns false at EOF") {
Review Comment:
I am getting error when running this test locally:
```
Caused by: java.util.NoSuchElementException
at org.apache.avro.file.DataFileStream.next(DataFileStream.java:262)
at org.apache.avro.file.DataFileStream.next(DataFileStream.java:248)
at org.apache.spark.sql.avro.AvroUtils$RowReader.hasNextRow(AvroUtils.scala:210)
at org.apache.spark.sql.avro.AvroUtils$RowReader.hasNextRow$(AvroUtils.scala:198)
at org.apache.spark.sql.avro.AvroFileFormat$$anon$1.hasNextRow(AvroFileFormat.scala:137)
```
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "gengliangwang (via GitHub)" <gi...@apache.org>.
gengliangwang commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529689335
##########
connector/avro/src/main/scala/org/apache/spark/sql/avro/AvroUtils.scala:
##########
@@ -197,19 +197,20 @@ private[sql] object AvroUtils extends Logging {
def hasNextRow: Boolean = {
while (!completed && currentRow.isEmpty) {
- if (fileReader.pastSync(stopPosition)) {
+ // In some cases of empty blocks in an Avro file, `fileReader.hasNext()` returns false but
Review Comment:
Could you explain what's the "some cases" here. It looks weird to call `hasNext` twice.
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "gengliangwang (via GitHub)" <gi...@apache.org>.
gengliangwang commented on PR #45578:
URL: https://github.com/apache/spark/pull/45578#issuecomment-2005803919
I think we can merge this fix first. And we can always revisit the code change here if Avro releases your fix. Thanks for the works!
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "cloud-fan (via GitHub)" <gi...@apache.org>.
cloud-fan commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529614907
##########
connector/avro/src/main/scala/org/apache/spark/sql/avro/AvroUtils.scala:
##########
@@ -197,19 +197,19 @@ private[sql] object AvroUtils extends Logging {
def hasNextRow: Boolean = {
while (!completed && currentRow.isEmpty) {
- if (fileReader.pastSync(stopPosition)) {
+ // In some cases of empty blocks in an Avro file, `fileReader.hasNext()` returns false but
+ // advances the cursor to the next block, so we need to call `fileReader.hasNext()` again
+ // to correctly report whether the next record exists.
+ val r = (fileReader.hasNext || fileReader.hasNext) && !fileReader.pastSync(stopPosition)
Review Comment:
```suggestion
val noMoreData = (fileReader.hasNext || fileReader.hasNext) && !fileReader.pastSync(stopPosition)
```
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "sadikovi (via GitHub)" <gi...@apache.org>.
sadikovi commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529706332
##########
connector/avro/src/test/scala/org/apache/spark/sql/avro/AvroSuite.scala:
##########
@@ -2744,6 +2744,16 @@ abstract class AvroSuite
}
}
}
+
+ test("SPARK-46990: read an empty file where pastSync returns false at EOF") {
Review Comment:
This could be related to your environment. The tests pass for me and on CI. Please share more details so I can debug further.
--
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-46990][SQL] Fix loading empty Avro files emitted by event-hubs [spark]
Posted by "sadikovi (via GitHub)" <gi...@apache.org>.
sadikovi commented on code in PR #45578:
URL: https://github.com/apache/spark/pull/45578#discussion_r1529737022
##########
connector/avro/src/main/scala/org/apache/spark/sql/avro/AvroUtils.scala:
##########
@@ -197,19 +197,20 @@ private[sql] object AvroUtils extends Logging {
def hasNextRow: Boolean = {
while (!completed && currentRow.isEmpty) {
- if (fileReader.pastSync(stopPosition)) {
+ // In some cases of empty blocks in an Avro file, `fileReader.hasNext()` returns false but
Review Comment:
It seems to be a bug in Avro. When blockRemaining can be 0, hasNext tries to load the next block but still checks if blockRemaining != 0 returning false when the next block is actually available.
The Avro FileReader API is limited and the only thing I could do is to just try to call hasNext again - seems to work for all of the tests cases including empty blocks.
You are right, ideally we should just loop over hasNext until it actually returns false or we reach EOF. I tried to implement it but I could not because FileReader does not expose the current stream offset (`tell()` actually returns the block start which is different).
--
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