Hi Micahm, thank you for your suggestions.

We have a sync discussing this in Parquet sync up meeting. Besides the meta 
data items suggested by you (compression level and block size), we will also 
provide information like vendor name, versions. Will refine the proposal 
further 
(https://docs.google.com/document/d/1ueSYq2FIzaom23cpHXppig93ylOxe8CU6EwS82dov2E/edit#heading=h.5b2qz2ba32wm
 for better access).

In-compatible with standard compression codec is not the expectation here. The 
expected behavior here is: 1) leveraging user defined or optimized 
implementation provided by users as first; 2) fallback to standard codec 
implementation. If compatible issue happens, as a framework, we will throw 
error message to end users directly.

Thanks
Cheng Xu

-----Original Message-----
From: Micah Kornfield <[email protected]> 
Sent: Tuesday, June 23, 2020 11:57 AM
To: Parquet Dev <[email protected]>
Subject: Re: Proposal for CompressionCodec Provider-aware Compression Codec 
Lookup for parquet-mr

Instead of a custom compressor name is there some way to expose more metadata 
about the parameters a particular codec used for compression (e.g.
compression level used or block size) be sufficient?  I'm not sure how 
standardized these are across given implementations/versions of the codecs 
currently supported so it might not be.

Just to clarify the proposal.  Is the suggestion that there will be some 
compression algorithms that won't be decodable by standard library 
implementation for each codec?  I agree with Gabor that this shouldn't be 
supported.

If the only requirement is, that in some cases, a customized decoder needs to 
be supported, and the customized decoder might not support decoding all encoded 
data then this seems less bad.  However it is still a slippery slope.  If this 
is the case another two options to consider:
1.  Always use the plugin by default and if it fails fallback to the normal 
codec.
2.  Don't specify a name specifically but add the customization hook to accept 
the entire set of metadata from the footer.

-Micah

-Micah

On Mon, Jun 22, 2020 at 6:56 AM Xu, Cheng A <[email protected]> wrote:

> Thanks Gabor for the comments.
> https://docs.google.com/document/d/1ueSYq2FIzaom23cpHXppig93ylOxe8CU6E
> wS82dov2E/edit
> Updated with comment access.
>
> Yes, ideally, we should have all codec backward compatible with 
> customized ones. However, in some cases, it's hard to support that. 
> For some users, they may reply on some accelerators to do the 
> compression work. Those accelerators are limited in memory which 
> doesn't allow a large history buffer to decompress.
>
> My understanding for this proposal is we try to introduce a framework 
> to allow customers customize their compression codec. And it's 
> customer's own responsibility if they use in-compatible format in 
> return with good performance.
> This is similar to what airlift did. Airlift is actually a codec provider.
> It provides a few codec supported by Parquet. We can have some 
> official supported codec provider IDs like built-in, airlift. And 
> users can make their own decisions to extend providers with their new codec 
> providers.
>
> Your thoughts on this?
>
> Thanks
> Cheng Xu
>
> -----Original Message-----
> From: Gabor Szadovszky <[email protected]>
> Sent: Monday, June 22, 2020 5:09 PM
> To: Parquet Dev <[email protected]>
> Subject: Re: Proposal for CompressionCodec Provider-aware Compression 
> Codec Lookup for parquet-mr
>
> Hi Cheng Xu,
>
> It would be easier if we would have comment access to the document.
> After the first look I have the following comments:
> - "different [codec] implementations may not be compatible with others 
> due to different purposes." - This is a huge problem. Parquet 
> specifies the compression codecs that the format supports. We've 
> already had issues by not specifying the codecs properly (see 
> PARQUET-1241 < https://issues.apache.org/jira/browse/PARQUET-1241> for 
> details). We shall not allow situations like this one. If a parquet 
> file is written with a compression codec from the spec shall be 
> readable by another parquet implementation that supports that codec 
> independently from the provider.
> - providers of the compression codecs are usually implementation dependent.
> How would different parquet implementations handle the different providers?
> (e.g. a java based compression provider is to be used by parquet-cpp)
> - how do we specify the provider names?
>
> Regards,
> Gabor
>
> On Fri, Jun 19, 2020 at 4:30 PM Xu, Cheng A <[email protected]> wrote:
>
> > Hi folks, any suggestions on this?
> >
> > Thanks
> > Cheng Xu
> >
> > -----Original Message-----
> > From: Dong, Xin <[email protected]>
> > Sent: Friday, June 5, 2020 2:19 PM
> > To: [email protected]
> > Subject: RE: Proposal for CompressionCodec Provider-aware 
> > Compression Codec Lookup for parquet-mr
> >
> > Hi, Walid,
> >
> > We've moved the doc here for public access:
> >
> > https://docs.google.com/document/d/1ueSYq2FIzaom23cpHXppig93ylOxe8CU
> > 6E
> > wS82dov2E/
> >
> > Thanks,
> > Xin Dong
> >
> > -----Original Message-----
> > From: Gara Walid <[email protected]>
> > Sent: Thursday, June 4, 2020 2:14 PM
> > To: [email protected]
> > Subject: Re: Proposal for CompressionCodec Provider-aware 
> > Compression Codec Lookup for parquet-mr
> >
> > Hi Xin,
> >
> > Thanks for the proposal. Could you please make the google doc public?
> >
> > Cheers,
> > Walid
> >
> > On Thu, Jun 4, 2020, 6:46 AM Dong, Xin <[email protected]> wrote:
> >
> > > Hi, All,
> > >
> > > The existing Parquet compress codec framework only supports codec 
> > > name based compression implementation lookup. And it's one-2-one 
> > > mapping which means only one implementation is supported given a 
> > > codec
> name.
> > > However, there are various implementations for the same codec name.
> > > And different implementations may not be compatible with others 
> > > due to different purposes. Given Gzip as an example, for some 
> > > accelerators, it's limited in memory capacity and the history 
> > > buffer size is relatively smaller than CPU based.  And currently 
> > > codec framework doesn't provide a mechanism to allow users to 
> > > customize standard compression codec for their own purposes (e.g. 
> > > performance acceleration,
> > workload offloading).
> > > To address the problem, we propose a provider-aware compression 
> > > codec lookup for parquet-mr. We've put the proposal here:
> > >
> > > https://docs.google.com/document/d/1sbCjDxEjM5UkbMPNmGqEfF-LYPDWhM
> > > -B
> > > 47 4dZZeOFD4/edit?ts=5ecb2462#heading=h.5b2qz2ba32wm
> > >
> > > Any comment is welcome and please let us know your feedback.
> > >
> > > Thanks,
> > > Xin Dong
> > >
> >
>

Reply via email to