Github user ramkrish86 commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#issuecomment-35471151
Hmmm..Got it in this context. Should have thought of this myself.. Thanks
James.
---
If your project is set up for it, you can reply to this email and have
Thanks for the patch, James. I like the idea - looks like JIRA is down at
the moment, so I can't take a look yet. Does the Apache commons-csv library
support custom delimiters? Any reason why the scripts can't be adapted to
the new library?
Thanks,
James
On Tue, Feb 18, 2014 at 2:47 PM, James Vi
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#issuecomment-35470690
We don't have any special cases for comparison of arrays. If we allow an
array to be part of the row key (or allow an array to be part of a secondary
index
Github user ramkrish86 commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#issuecomment-35469802
Hmm, makes sense, but what i felt is in case of comparisons of such arrays,
wont we specify the offset from where the comparison should happen? Leaving
out t
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#issuecomment-35469351
It's not the minor perf gain of comparing the header information that I'm
talking about. It's the fact that it makes comparing the byte arrays directly
impo
Github user ramkrish86 commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#issuecomment-35469062
Okie, so you mean avoiding the comparison of the first few bytes like
version, no of elements can be avoided.
---
If your project is set up for it, you can
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854998
You'd have to have the exact length specified which is nearly impossible
(especially for variable length). Just do the above and you'll prevent the copy
in al
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854967
You do still need to do some of this work, but you do it up front at
construction time. You just need to figure out the array element type first. It
can be ba
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#issuecomment-35468254
By comparison being easier, I mean you can compare the byte array directly
without deserializing it into objects, if the array type matches. The reason is
t
Github user ramkrish86 commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854849
Okie, so here on the ptr you mean. I foudn that internally if we are able
to specify a correct size for the TrustedByteStream then calling toByteArray
does n
Github user ramkrish86 commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854798
this is where i have a question. Now this serialization format should
comply with the one in the ParrayDataType where the normal array serialization
occurs and
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854769
This make a copy of the byte array which is bad. Do this instead:
ptr.get(byteStream.getBuffer(),0,byteStream.getSize());
---
If your project is
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854745
IOException
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. To do so, please top-post your res
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854741
Should only catch IOException, which is impossible for
TrustedByteArrayOutputStream
---
If your project is set up for it, you can reply to this email and hav
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854730
This is what you want to avoid. Just serialize the bytes into the oStream
and then write the offset information afterwards.
---
If your project is set up for
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854712
In this loop, you want to keep the offset info for variable length arrays.
Allocate a int array of size children.size() to put them in. You'll write them
all
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854686
You don't need this copy. Just do oStream.write(ptr.get(), ptr.getOffset(),
ptr.getLenth());
---
If your project is set up for it, you can reply to this emai
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854650
Change estimateSize to give a reasonable estimate for variable length
types. I believe we have a PDataType.estimateSize() that you can use.
---
If your proje
Github user JamesRTaylor commented on the pull request:
https://github.com/apache/incubator-phoenix/pull/8#discussion_r9854619
You should know already the type of the array and array element, so no need
to loop over the children to figure that out in the evaluate. The child type is
de
[
https://issues.apache.org/jira/browse/PHOENIX-56?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904877#comment-13904877
]
Russell Jurney edited comment on PHOENIX-56 at 2/19/14 12:41 AM:
---
[
https://issues.apache.org/jira/browse/PHOENIX-56?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904877#comment-13904877
]
Russell Jurney commented on PHOENIX-56:
---
James: we ran your class. Here is what it d
[
https://issues.apache.org/jira/browse/PHOENIX-15?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904856#comment-13904856
]
James Taylor commented on PHOENIX-15:
-
Thanks for finding *another* bug. Given that no
[
https://issues.apache.org/jira/browse/PHOENIX-58?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jesse Yates updated PHOENIX-58:
---
Attachment: phoenix-58-v0.patch
client-without-hbase-jar.txt
client-mi
Jesse Yates created PHOENIX-58:
--
Summary: Client jars are missing phoenix code
Key: PHOENIX-58
URL: https://issues.apache.org/jira/browse/PHOENIX-58
Project: Phoenix
Issue Type: Bug
Affects
[
https://issues.apache.org/jira/browse/PHOENIX-53?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
James Violette updated PHOENIX-53:
--
Attachment: incubator-phoenix-csv-commons.patch
This file incorporates the Apache commons-csv l
[
https://issues.apache.org/jira/browse/PHOENIX-53?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904732#comment-13904732
]
James Violette commented on PHOENIX-53:
---
We found that the apache commons-csv librar
[
https://issues.apache.org/jira/browse/PHOENIX-56?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904723#comment-13904723
]
Russell Jurney commented on PHOENIX-56:
---
James: Where does that class, PhoenixSqllin
[
https://issues.apache.org/jira/browse/PHOENIX-15?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lars Hofhansl updated PHOENIX-15:
-
Attachment: phoenix-15-v3.txt
Updated patch.
Note that there was a recent change to how the disab
[
https://issues.apache.org/jira/browse/PHOENIX-56?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Russell Jurney updated PHOENIX-56:
--
Attachment: hbase-site.xml
hbase-site.xml is attached, its the same for client and server.
> P
[
https://issues.apache.org/jira/browse/PHOENIX-56?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904668#comment-13904668
]
Russell Jurney commented on PHOENIX-56:
---
Zookeeper runs on the same nodes as HBase,
[
https://issues.apache.org/jira/browse/PHOENIX-57?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904632#comment-13904632
]
James Taylor commented on PHOENIX-57:
-
+1. Thanks for the fix, [~jesse_yates]
> Clien
[
https://issues.apache.org/jira/browse/PHOENIX-56?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904503#comment-13904503
]
Mujtaba Chohan commented on PHOENIX-56:
---
@russell.jurney you said that you are speci
[
https://issues.apache.org/jira/browse/PHOENIX-38?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904491#comment-13904491
]
Jesse Yates commented on PHOENIX-38:
awesome, thanks Mujtaba!
> Data load time with m
[
https://issues.apache.org/jira/browse/PHOENIX-38?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904485#comment-13904485
]
Mujtaba Chohan commented on PHOENIX-38:
---
Verified. Immutable index data load perform
[
https://issues.apache.org/jira/browse/PHOENIX-57?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904479#comment-13904479
]
Jesse Yates commented on PHOENIX-57:
[~giacomotaylor] mind taking a look? its a pretty
[
https://issues.apache.org/jira/browse/PHOENIX-57?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jesse Yates updated PHOENIX-57:
---
Attachment: phoenix-57-v0.patch
> ClientKeyValue doesn't work in debug mode
> ---
[
https://issues.apache.org/jira/browse/PHOENIX-57?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jesse Yates reassigned PHOENIX-57:
--
Assignee: Jesse Yates
> ClientKeyValue doesn't work in debug mode
> ---
Jesse Yates created PHOENIX-57:
--
Summary: ClientKeyValue doesn't work in debug mode
Key: PHOENIX-57
URL: https://issues.apache.org/jira/browse/PHOENIX-57
Project: Phoenix
Issue Type: Bug
Aff
[
https://issues.apache.org/jira/browse/PHOENIX-57?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904476#comment-13904476
]
Jesse Yates commented on PHOENIX-57:
Since this only occurs in debugging, we don't nee
GitHub user ramkrish86 opened a pull request:
https://github.com/apache/incubator-phoenix/pull/8
Phoenix 37
First cut pull request. Still not changed the format of serialization of
variable length array in ArrayConstructorExpression.
Moved the HEader information of the array'
[
https://issues.apache.org/jira/browse/PHOENIX-56?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904380#comment-13904380
]
Mujtaba Chohan commented on PHOENIX-56:
---
Sure [~jamestaylor] I'll try to repro as we
[
https://issues.apache.org/jira/browse/PHOENIX-56?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904268#comment-13904268
]
James Taylor commented on PHOENIX-56:
-
[~mujtabachohan] - would you have time to try t
[
https://issues.apache.org/jira/browse/PHOENIX-37?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904240#comment-13904240
]
James Taylor commented on PHOENIX-37:
-
Use org.apache.phoenix.util.TrustedByteArrayOut
[
https://issues.apache.org/jira/browse/PHOENIX-21?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eli Levine updated PHOENIX-21:
--
Comment: was deleted
(was: I can envision an SFDC scenario with tenants having thousands of
tenant-spe
[
https://issues.apache.org/jira/browse/PHOENIX-21?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904225#comment-13904225
]
Eli Levine commented on PHOENIX-21:
---
Looks good. Thanks!
> Support indexes on multi-te
[
https://issues.apache.org/jira/browse/PHOENIX-37?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13904024#comment-13904024
]
ramkrishna.s.vasudevan commented on PHOENIX-37:
---
So for the variable length
[
https://issues.apache.org/jira/browse/PHOENIX-21?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13903959#comment-13903959
]
Hudson commented on PHOENIX-21:
---
SUCCESS: Integrated in Apache Phoenix - Branch:master #55 (
[
https://issues.apache.org/jira/browse/PHOENIX-21?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
James Taylor updated PHOENIX-21:
Attachment: tenant-specific-index-2.patch
> Support indexes on multi-tenant views
> ---
[
https://issues.apache.org/jira/browse/PHOENIX-21?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
James Taylor updated PHOENIX-21:
Attachment: (was: 7b.patch)
> Support indexes on multi-tenant views
> -
[
https://issues.apache.org/jira/browse/PHOENIX-21?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13903903#comment-13903903
]
James Taylor commented on PHOENIX-21:
-
I've made it so that non-tenant specific indexe
[
https://issues.apache.org/jira/browse/PHOENIX-21?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
James Taylor updated PHOENIX-21:
Attachment: 7b.patch
> Support indexes on multi-tenant views
>
51 matches
Mail list logo
