After getting in PARQUET-1625 I vote again for having bloom filter spec and the thrift file update as is in parquet-format master. +1 (binding)
On Mon, Jul 15, 2019 at 3:23 PM 俊杰陈 <[email protected]> wrote: > Thanks Gabor, It's never too late to make it better. We don't have to > run it in a hurry, it has been developed for a long time yet.:) > > The thrift file is indeed a bit lag behind the spec. As the spec > defined, the bloom filter data is stored near the footer which means > we don't have to handle it like the page. Therefore, I just opened a > jira to remove bloom_filter_page_header in PageHeader structure, while > the BloomFitlerHeader is kept intentionally for convenience. Since the > spec and the thrift should be aligned with each other eventually, so > the vote target is both of them. > > > > On Mon, Jul 15, 2019 at 7:48 PM Gabor Szadovszky > <[email protected]> wrote: > > > > Hi Junjie, > > > > Sorry for bringing up this a bit late but I have some problems with the > > format update. The parquet.thrift file is updated to have the bloom > filters > > as a page (just as dictionaries and data pages). Meanwhile, the spec > > (BloomFilter.md) says that the bloom filter is stored near the footer. > So, > > if the bloom filter is not part of the row-groups (like column indexes) I > > would not add it as a page. See the struct ColumnIndex in the thrift > file. > > This struct is not referenced anywhere in it only declared. It was done > > this way because we don't parse it in the same way as we parse the pages. > > > > Currently, I am not 100% sure about the target of this vote. If it is a > > vote about adding bloom filters in general then it is a +1 (binding). If > it > > is about adding the bloom filters to parquet-format as is then, it is a > -1 > > (binding) until we fix the issue above. > > > > Regards, > > Gabor > > > > On Mon, Jul 15, 2019 at 11:45 AM Gidon Gershinsky <[email protected]> > wrote: > > > > > +1 (non-binding) > > > > > > On Mon, Jul 15, 2019 at 12:08 PM Zoltan Ivanfi <[email protected] > > > > > wrote: > > > > > > > +1 (binding) > > > > > > > > On Mon, Jul 15, 2019 at 9:57 AM 俊杰陈 <[email protected]> wrote: > > > > > > > > > > Dear Parquet developers > > > > > > > > > > I'd like to resume this vote, you can start to vote now. Thanks for > > > your > > > > time. > > > > > > > > > > On Wed, Jul 10, 2019 at 9:29 PM 俊杰陈 <[email protected]> wrote: > > > > > > > > > > > > I see, will resume this next week. Thanks. > > > > > > > > > > > > > > > > > > > > > > > > On Wed, Jul 10, 2019 at 5:26 PM Zoltan Ivanfi > > > <[email protected]> > > > > wrote: > > > > > > > > > > > > > > Hi Junjie, > > > > > > > > > > > > > > Since there are ongoing improvements addressing review > comments, I > > > > would > > > > > > > hold off with the vote for a few more days until the > specification > > > > settles. > > > > > > > > > > > > > > Br, > > > > > > > > > > > > > > Zoltan > > > > > > > > > > > > > > On Wed, Jul 10, 2019 at 9:32 AM 俊杰陈 <[email protected]> > wrote: > > > > > > > > > > > > > > > Hi Parquet committers and developers > > > > > > > > > > > > > > > > We are waiting for your important ballot:) > > > > > > > > > > > > > > > > On Tue, Jul 9, 2019 at 10:21 AM 俊杰陈 <[email protected]> > wrote: > > > > > > > > > > > > > > > > > > Yes, there are some public benchmark results, such as the > > > > official > > > > > > > > > benchmark from xxhash site (http://www.xxhash.com/) and > > > > published > > > > > > > > > comparison from smhasher project > > > > > > > > > (https://github.com/rurban/smhasher/). > > > > > > > > > > > > > > > > > > > > > > > > > > > On Tue, Jul 9, 2019 at 5:25 AM Wes McKinney < > > > [email protected]> > > > > wrote: > > > > > > > > > > > > > > > > > > > > Do you have any benchmark data to support the choice of > hash > > > > function? > > > > > > > > > > > > > > > > > > > > On Wed, Jul 3, 2019 at 8:41 AM 俊杰陈 <[email protected]> > > > wrote: > > > > > > > > > > > > > > > > > > > > > > Dear Parquet developers > > > > > > > > > > > > > > > > > > > > > > To simplify the voting, I 'd like to update voting > content > > > > to the > > > > > > > > spec > > > > > > > > > > > with xxHash hash strategy. Now you can reply with +1 > or -1. > > > > > > > > > > > > > > > > > > > > > > Thanks for your participation. > > > > > > > > > > > > > > > > > > > > > > On Tue, Jul 2, 2019 at 10:23 AM 俊杰陈 < > [email protected]> > > > > wrote: > > > > > > > > > > > > > > > > > > > > > > > > Dear Parquet developers > > > > > > > > > > > > > > > > > > > > > > > > Parquet Bloom filter has been developed for a while, > per > > > > the > > > > > > > > discussion on the mail list, it's time to call a vote for > spec to > > > > move > > > > > > > > forward. The current spec can be found at > > > > > > > > > > > > https://github.com/apache/parquet-format/blob/master/BloomFilter.md. > > > > > > > > There are some different options about the internal hash > choice > > > of > > > > Bloom > > > > > > > > filter and the PR is for that concern. > > > > > > > > > > > > > > > > > > > > > > > > So I 'd like to propose to vote the spec + hash > option, > > > for > > > > > > > > example: > > > > > > > > > > > > > > > > > > > > > > > > +1 to spec and xxHash > > > > > > > > > > > > +1 to spec and murmur3 > > > > > > > > > > > > ... > > > > > > > > > > > > > > > > > > > > > > > > Please help to vote, any feedback is also welcome in > the > > > > > > > > discussion thread. > > > > > > > > > > > > > > > > > > > > > > > > Thanks & Best Regards > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > > > > > Thanks & Best Regards > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > > > Thanks & Best Regards > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > > Thanks & Best Regards > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > Thanks & Best Regards > > > > > > > > > > > > > > > > > > > > -- > > > > > Thanks & Best Regards > > > > > > > > > > > -- > Thanks & Best Regards >
