You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@arrow.apache.org by GitBox <gi...@apache.org> on 2022/04/19 13:02:40 UTC
[GitHub] [arrow-datafusion] thinkharderdev opened a new pull request, #2273: Add bytes scanned metric to ParquetExec
thinkharderdev opened a new pull request, #2273:
URL: https://github.com/apache/arrow-datafusion/pull/2273
# Which issue does this PR close?
<!--
We generally require a GitHub issue to be filed for all bug fixes and enhancements and this helps us generate change logs for our releases. You can link an issue to this PR using the GitHub syntax. For example `Closes #123` indicates that this PR will close issue #123.
-->
Closes #.
# Rationale for this change
<!--
Why are you proposing this change? If this is already explained clearly in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand your changes and offer better suggestions for fixes.
-->
Tracking total number of bytes read from storage is an important metric for query execution (especially for use cases where data is read from object storage or NAS rather than a local SSD). It's especially important for Parquet data sources where we have many levers we can pull to reduce the amount of data read.
# What changes are included in this PR?
<!--
There is no need to duplicate the description in the issue here but it is sometimes worth providing a summary of the individual changes in this PR.
-->
Add a `bytes_scanned` counter to `ParquetFileMetrics` and track total number of bytes read through `ChunkObjectReader`. This should be quite cheap since we pass the number of bytes to read as an argument.
# Are there any user-facing changes?
<!--
If there are user-facing changes then we may require documentation to be updated before approving the PR.
-->
`ParquetExec` will have a `bytes_scanned` metric.
<!--
If there are any breaking changes to public APIs, please add the `api change` label.
-->
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] andygrove commented on a diff in pull request #2273: Add bytes scanned metric to ParquetExec
Posted by GitBox <gi...@apache.org>.
andygrove commented on code in PR #2273:
URL: https://github.com/apache/arrow-datafusion/pull/2273#discussion_r854317529
##########
datafusion/core/src/datasource/file_format/parquet.rs:
##########
@@ -362,19 +368,27 @@ fn fetch_statistics(
}
/// A wrapper around the object reader to make it implement `ChunkReader`
-pub struct ChunkObjectReader(pub Arc<dyn ObjectReader>);
+pub struct ChunkObjectReader {
+ /// The underlying object reader
+ pub object_reader: Arc<dyn ObjectReader>,
+ /// Optional counter which will track total number of bytes scanned
+ pub bytes_scanned: Option<metrics::Count>,
+}
impl Length for ChunkObjectReader {
fn len(&self) -> u64 {
- self.0.length()
+ self.object_reader.length()
}
}
impl ChunkReader for ChunkObjectReader {
type T = Box<dyn Read + Send + Sync>;
fn get_read(&self, start: u64, length: usize) -> ParquetResult<Self::T> {
- self.0
+ if let Some(m) = self.bytes_scanned.as_ref() {
+ m.add(length)
Review Comment:
I'm just curious .. the length here is the number of bytes requested and could potentially be different from the actual number of bytes read?
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] thinkharderdev commented on a diff in pull request #2273: Add bytes scanned metric to ParquetExec
Posted by GitBox <gi...@apache.org>.
thinkharderdev commented on code in PR #2273:
URL: https://github.com/apache/arrow-datafusion/pull/2273#discussion_r854461112
##########
datafusion/core/src/datasource/file_format/parquet.rs:
##########
@@ -362,19 +368,27 @@ fn fetch_statistics(
}
/// A wrapper around the object reader to make it implement `ChunkReader`
-pub struct ChunkObjectReader(pub Arc<dyn ObjectReader>);
+pub struct ChunkObjectReader {
+ /// The underlying object reader
+ pub object_reader: Arc<dyn ObjectReader>,
+ /// Optional counter which will track total number of bytes scanned
+ pub bytes_scanned: Option<metrics::Count>,
+}
impl Length for ChunkObjectReader {
fn len(&self) -> u64 {
- self.0.length()
+ self.object_reader.length()
}
}
impl ChunkReader for ChunkObjectReader {
type T = Box<dyn Read + Send + Sync>;
fn get_read(&self, start: u64, length: usize) -> ParquetResult<Self::T> {
- self.0
+ if let Some(m) = self.bytes_scanned.as_ref() {
+ m.add(length)
Review Comment:
I don't think this interface actually specifies so it would really be up to the specific `ObjectStore` implementation. I know that that is the way the `LocalFileSystem` works and I believe that is the way the AWS S3 API works with file ranges.
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] andygrove merged pull request #2273: Add bytes scanned metric to ParquetExec
Posted by GitBox <gi...@apache.org>.
andygrove merged PR #2273:
URL: https://github.com/apache/arrow-datafusion/pull/2273
--
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: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org