-
Notifications
You must be signed in to change notification settings - Fork 232
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Support int96RebaseModeInWrite
and int96RebaseModeInRead
#3330
Conversation
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
build |
int96RebaseModeInWrite
and int96RebaeModeInRead
int96RebaseModeInWrite
and int96RebaseModeInRead
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I was expecting a shim function that would check INT96 timestamp rebase node, and in Spark < 3.1.1 it would check the old datetime config and in Spark >= 3.1.1 it would check the new INT96-specific configs.
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetFileFormat.scala
Outdated
Show resolved
Hide resolved
Ironically, that was my first implementation. The reason why I didn't like that was because we will be returning the value of |
But that is exactly what Spark < 3.1.1 is doing today. The shim function name would indicate we're trying to get the int96 rebase mode, and on Spark >= 3.1.1 it would check the Int96-specific config, and on Spark < 3.1.1 it would check the same config the Spark code is checking in that situation. |
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
I have made some unnecessary changes to the SparkBaseShim in every version of 311. Will revert in a bit |
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
build |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The Databricks shims will also need to be updated accordingly.
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetFileFormat.scala
Outdated
Show resolved
Hide resolved
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
build |
NFS issue, reported to SRE |
build |
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetFileFormat.scala
Outdated
Show resolved
Hide resolved
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetFileFormat.scala
Outdated
Show resolved
Hide resolved
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetFileFormat.scala
Outdated
Show resolved
Hide resolved
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetScan.scala
Outdated
Show resolved
Hide resolved
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetScan.scala
Outdated
Show resolved
Hide resolved
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetScan.scala
Outdated
Show resolved
Hide resolved
sql-plugin/src/main/scala/com/nvidia/spark/rapids/GpuParquetScan.scala
Outdated
Show resolved
Hide resolved
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
build |
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
build |
1 similar comment
build |
@razajafri hi, I saw you triggered premerge twice but not abort the previous build, may I ask why? each build would take 2GPUs |
Need to resolve merge conflicts but otherwise lgtm. |
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
build |
Looks like a failure unrelated to my change in DB, root causing it |
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
Skipping |
build |
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
This reverts commit 14658b4. Signed-off-by: Raza Jafri <rjafri@nvidia.com>
Signed-off-by: Raza Jafri <rjafri@nvidia.com>
build |
Co-authored-by: Jason Lowe <jlowe@nvidia.com>
build |
int96RebaseModeInWrite
and int96RebaseModeInRead
[databricks]int96RebaseModeInWrite
and int96RebaseModeInRead
…VIDIA#3330)" This reverts commit fc40c00.
this pr is not up to date with latest moves of of shim files, reverted #3627 |
While writing/reading a parquet file a user can specify a different config for reading/writing int96 values vs datetime values in general.
This PR takes the value set by users to
int96RebaseModeInWrite
andint96RebaseModeInRead
and matches what Apache spark does.Signed-off-by: Raza Jafri rjafri@nvidia.com