[SPARK-31885][SQL][3.0] Fix filter push down for old millis timestamps to Parquet #28699
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…Parquet
Fixed conversions of `java.sql.Timestamp` to milliseconds in `ParquetFilter` by using existing functions from `DateTimeUtils` `fromJavaTimestamp()` and `microsToMillis()`.
The changes fix the bug:
```scala
scala> spark.conf.set("spark.sql.parquet.outputTimestampType", "TIMESTAMP_MILLIS")
scala> spark.conf.set("spark.sql.legacy.parquet.datetimeRebaseModeInWrite", "CORRECTED")
scala> Seq(java.sql.Timestamp.valueOf("1000-06-14 08:28:53.123")).toDF("ts").write.mode("overwrite").parquet("/Users/maximgekk/tmp/ts_millis_old_filter")
scala> spark.read.parquet("/Users/maximgekk/tmp/ts_millis_old_filter").filter($"ts" === "1000-06-14 08:28:53.123").show(false)
+---+
|ts |
+---+
+---+
```
Yes, after the changes (for the example above):
```scala
scala> spark.read.parquet("/Users/maximgekk/tmp/ts_millis_old_filter").filter($"ts" === "1000-06-14 08:28:53.123").show(false)
+-----------------------+
|ts |
+-----------------------+
|1000-06-14 08:28:53.123|
+-----------------------+
```
Modified tests in `ParquetFilterSuite` to check old timestamps.
Closes apache#28693 from MaxGekk/parquet-ts-millis-filter.
Authored-by: Max Gekk <[email protected]>
Signed-off-by: Wenchen Fan <[email protected]>
(cherry picked from commit 9c0dc28)
Signed-off-by: Max Gekk <[email protected]>
MaxGekk
commented
Jun 1, 2020
| private def timestampToMillis(v: Any): JLong = { | ||
| val timestamp = v.asInstanceOf[Timestamp] | ||
| val micros = DateTimeUtils.fromJavaTimestamp(timestamp) | ||
| val millis = DateTimeUtils.toMillis(micros) |
Member
Author
There was a problem hiding this comment.
This line causes the problem.
|
Test build #123390 has finished for PR 28699 at commit
|
Member
Author
|
jenkins, retest this, please |
HyukjinKwon
approved these changes
Jun 2, 2020
Contributor
|
retest this please |
|
Test build #123397 has finished for PR 28699 at commit
|
Contributor
|
thanks, merging to 3.0! |
cloud-fan
pushed a commit
that referenced
this pull request
Jun 2, 2020
…s to Parquet
### What changes were proposed in this pull request?
Fixed conversions of `java.sql.Timestamp` to milliseconds in `ParquetFilter` by using existing functions from `DateTimeUtils` `fromJavaTimestamp()` and `microsToMillis()`.
### Why are the changes needed?
The changes fix the bug:
```scala
scala> spark.conf.set("spark.sql.parquet.outputTimestampType", "TIMESTAMP_MILLIS")
scala> spark.conf.set("spark.sql.legacy.parquet.datetimeRebaseModeInWrite", "CORRECTED")
scala> Seq(java.sql.Timestamp.valueOf("1000-06-14 08:28:53.123")).toDF("ts").write.mode("overwrite").parquet("/Users/maximgekk/tmp/ts_millis_old_filter")
scala> spark.read.parquet("/Users/maximgekk/tmp/ts_millis_old_filter").filter($"ts" === "1000-06-14 08:28:53.123").show(false)
+---+
|ts |
+---+
+---+
```
### Does this PR introduce _any_ user-facing change?
Yes, after the changes (for the example above):
```scala
scala> spark.read.parquet("/Users/maximgekk/tmp/ts_millis_old_filter").filter($"ts" === "1000-06-14 08:28:53.123").show(false)
+-----------------------+
|ts |
+-----------------------+
|1000-06-14 08:28:53.123|
+-----------------------+
```
### How was this patch tested?
Modified tests in `ParquetFilterSuite` to check old timestamps.
Closes #28699 from MaxGekk/parquet-ts-millis-filter-3.0.
Authored-by: Max Gekk <[email protected]>
Signed-off-by: Wenchen Fan <[email protected]>
|
Test build #123411 has finished for PR 28699 at commit
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
Fixed conversions of
java.sql.Timestampto milliseconds inParquetFilterby using existing functions fromDateTimeUtilsfromJavaTimestamp()andmicrosToMillis().Why are the changes needed?
The changes fix the bug:
Does this PR introduce any user-facing change?
Yes, after the changes (for the example above):
How was this patch tested?
Modified tests in
ParquetFilterSuiteto check old timestamps.