This is an automated email from the ASF dual-hosted git repository.
dzamo pushed a commit to branch gh-pages
in repository https://gitbox.apache.org/repos/asf/drill.git
The following commit(s) were added to refs/heads/gh-pages by this push:
new b9a1d50 Update 024-aynchronous-parquet-reader.md (#2280)
b9a1d50 is described below
commit b9a1d5095abeb6a538cb0e0b8a3003369f0d9dbf
Author: Mark Rajcok <[email protected]>
AuthorDate: Sun Jul 25 01:20:58 2021 -0400
Update 024-aynchronous-parquet-reader.md (#2280)
default pagereader.buffersize value was changed from 4 MiB to 1 MiB back in
2017:
https://github.com/apache/drill/commit/052010108a47856f9b1a3c0c470b6572948dc749#diff-72db537e92cd35c02f9e30e13e11421326ef03a05ebce3c22b29dd44d7893262R179.
See also
https://github.com/apache/drill/commit/a51c98b8bf210bbe9d3f4018361d937252d1226d#diff-10939a48c16faa36c4f0223e4b62012b8518f884e5b09bf5c88c3735865993eeR486
---
_docs/en/performance-tuning/024-aynchronous-parquet-reader.md | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/_docs/en/performance-tuning/024-aynchronous-parquet-reader.md
b/_docs/en/performance-tuning/024-aynchronous-parquet-reader.md
index 347597c..6a26c53 100644
--- a/_docs/en/performance-tuning/024-aynchronous-parquet-reader.md
+++ b/_docs/en/performance-tuning/024-aynchronous-parquet-reader.md
@@ -29,7 +29,7 @@ The following table lists and describes the asynchronous
Parquet reader options
|----------------------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|---------|-------------|
| store.parquet.reader.pagereader.async | Enable the asynchronous page
reader. This pipelines the reading of data from disk for high performance.
| BOOLEAN | TRUE |
| store.parquet.reader.pagereader.bufferedread | Enable buffered page reading.
Can improve disk scan speeds by buffering data, but increases memory usage.
This option is less useful when the number of columns increases.
| BOOLEAN | TRUE |
-| store.parquet.reader.pagereader.buffersize | The size of the buffer (in
bytes) to use if bufferedread is true. Has no effect otherwise.
| LONG | 4194304 |
+| store.parquet.reader.pagereader.buffersize | The size of the buffer (in
bytes) to use if bufferedread is true. Has no effect otherwise.
| LONG | 1048576 |
| store.parquet.reader.pagereader.usefadvise | If the file system supports
it, the Parquet file reader issues an fadvise call to enable file server side
sequential reading and caching. Since many HDFS implementations do not
support this and because this may have no effect in conditions of high
concurrency, the option is set to false. Useful for benchmarks and for
performance critical queries. | BOOLEAN | FALSE |
| store.parquet.reader.columnreader.async | Turn on parallel decoding of
column data from Parquet to the in memory format. This increases CPU usage
and is most useful for compressed fixed width data. With increasing
concurrency, this option may cause queries to run slower and should be turned
on only for performance critical queries.
| BOOLEAN | FALSE |
