PARQUET-521: Add Brotli compression codec. #344
Closed
Conversation
This file contains 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
| protected int getMaxCompressedLength(int numInputBytes) { | ||
| // this is based on https://github.com/google/brotli/issues/274 | ||
| // that page is not very clear, so this is much more conservative | ||
| return numInputBytes * 2; |
There was a problem hiding this comment.
seems like one of the linked issues pointed to this in the Brotli src: https://github.com/google/brotli/search?utf8=%E2%9C%93&q=max_out_size
const size_t max_out_size = 2 * bytes + 500;
There was a problem hiding this comment.
There's also a Brotli method that gives the max compressed length that we should get into the JNI bindings. This value has been working fine for all of my tests, but I'll update the value on the next pass.
rdblue
force-pushed
the
PARQUET-521-add-brotli-compression
branch
from
bea6774
to
28c52d1
Compare
This adds a Brotli codec that shares code with the Snappy codec. Snappy is "non-blocking", meaning that it always accepts more data and buffers it without blocking. The first reads is blocking, while the compression is done off-heap. This strategy doesn't appear to have a noticable performance impact, but does get better compression for Brotli than streaming buffers. The non-blocking part of Parquet's Snappy codec has been refactored so it can be used for both Brotli and Snappy, which both use direct byte buffers and compress outside the JVM.
rdblue
force-pushed
the
PARQUET-521-add-brotli-compression
branch
from
28c52d1
to
5938fb5
Compare
|
This is replaced by #430. Brotli is provided by https://github.com/rdblue/brotli-codec. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This adds a Brotli codec that shares code with the Snappy codec. Snappy
is "non-blocking", meaning that it always accepts more data and buffers
it without blocking. The first reads is blocking, while the compression
is done off-heap. This strategy doesn't appear to have a noticable
performance impact, but does get better compression for Brotli than
streaming buffers.
The non-blocking part of Parquet's Snappy codec has been refactored so
it can be used for both Brotli and Snappy, which both use direct byte
buffers and compress outside the JVM.
This currently depends on a snapshot release of parquet-format with
PARQUET-609.