Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Thu, 13 Oct 2016 16:35:20 +0000 (UTC)
From: "ramkrishna.s.vasudevan (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.13008481.1475141461000.818772.1476376520841@Atlassian.JIRA>
In-Reply-To: <JIRA.13008481.1475141461000@Atlassian.JIRA>
References: <JIRA.13008481.1475141461000@Atlassian.JIRA> <JIRA.13008481.1475141461714@arcas>
Subject: [jira] [Commented] (HBASE-16729) Define the behavior of (default)
 empty FilterList
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
archived-at: Thu, 13 Oct 2016 16:35:27 -0000


    [ https://issues.apache.org/jira/browse/HBASE-16729?page=3Dcom.atlassia=
n.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=3D155=
72437#comment-15572437 ]=20

ramkrishna.s.vasudevan commented on HBASE-16729:
------------------------------------------------

Ping [~saint.ack@gmail.com] and [~lhofhansl]. What do you think with respec=
t to this comment?
bq.Instead of any change, we can just add enough comments to explain the em=
pty FilterList. It seems to me that the issue of empty FilterList is the in=
consistent result(see HBASE-16731).
Do you reckon marking this as incompatible change as done in this patch or =
just a doc is enough? Seems [~chia7712] has run into issues because of empt=
y filter list. So I felt changing the behaviour and having the write behavi=
our is better than just doc update.

> Define the behavior of (default) empty FilterList
> -------------------------------------------------
>
>                 Key: HBASE-16729
>                 URL: https://issues.apache.org/jira/browse/HBASE-16729
>             Project: HBase
>          Issue Type: Wish
>            Reporter: ChiaPing Tsai
>            Assignee: ChiaPing Tsai
>            Priority: Trivial
>         Attachments: HBASE-16729.v0.patch
>
>
> Current empty FilterList filters all data, because the FilterList#isFamil=
yEssential always returns false which causes the null cell retrieved by Reg=
ionScannerImpl.storeHeap.
> It seems to me that empty FilterList should do nothing, because the follo=
wing code is common.
> {noformat}
> private static Filter makeFilter() {
> =C2=A0 FilterList filterList =3D new FilterList ();
> =C2=A0 for (some conditions) {
> =C2=A0 =C2=A0 // add some filters. Or nothing to add.
> =C2=A0 }
> =C2=A0 return filterList;
> }
> {noformat}
> If we keep the current logic which filters all data, we should add enough=
 comments to explain it. Or add the FilterList#size() or FilterList#empty()=
 for preventing filtering all data.
> Any comments? Thanks.


--
This message was sent by Atlassian JIRA
(v6.3.4#6332)