Couple of follow up comments:
- if you use c1=‘X0’ the index should be used without a hint, because it’s
still ordered by the PK when using index.
- this wouldn’t necessarily be the case for c1 LIKE 'X0%'.
On Fri, Oct 13, 2017 at 8:33 PM James Taylor <jamestay...@apache.org> wrote:
Yes, this is expected behavior. Phoenix can either optimize based on the
filter in the WHERE clause or the ORDER BY. Since it's not cost based,
Phoenix always chooses to optimize out the ORDER BY (because in general
it's more expensive and it doesn't know how much data will be filtered out
by the
The Apache Phoenix team is pleased to announce the immediate availability
of the 4.12.0 release [1]. Apache Phoenix enables SQL-based OLTP and
operational analytics for Apache Hadoop using Apache HBase as its backing
store and providing integration with other projects in the Apache ecosystem
such
gle.ca/amp/blog.colinbreck.com/calling-blocki
>> ng-code-there-is-no-free-lunch/amp/
>>
>> On Wed, Oct 4, 2017 at 11:36 AM Stan Campbell <stan.campbe...@gmail.com>
>> wrote:
>>
>>> Wrap the call in a Future. You're home.
>>>
>>> On Wed, Oc
Hi Flavio,
Phoenix supports JDBC. The implementation may do gets, scans, etc., but
it's completely transparent to the user.
Thanks,
James
On Wed, Oct 4, 2017 at 6:36 AM, Flavio Pompermaier
wrote:
> Hi to all,
> does Phoenix support async get? I can't find anything about
rget-temp table (after firing all queries) , and
>drop temp-table
>
> though this approach results in firing more queries, it can complete in
> less time than firing queries without using index
>
> On Wed, Sep 20, 2017 at 9:31 PM, James Taylor <jamestay...@apache.org>
>
Ryan - did the work around solve the issue?
Thanks,
James
On Fri, Sep 22, 2017 at 11:22 AM, Maryann Xue wrote:
> Hi Ryan,
>
> Could you please try:
>
> select
>
> MAIN.EBELN as PO_DOC,
>
> MAIN.EBELP as PO_DOC_ITEM_NUM,
>
> XREF.DELV_DOC
>
> from EKPO as MAIN
>
> left
Great to have you on the PMC, Sergey. Congrats!
James
On Mon, Sep 25, 2017 at 10:28 AM, Josh Mahonin wrote:
> Congratulations Sergey!
>
> On Sun, Sep 24, 2017 at 4:05 PM, Ted Yu wrote:
>
> > Congratulations, Sergey !
> >
> > On Sun, Sep 24, 2017 at
tiple indexes?
> Hint on local index cause an error of unknown field if the where clause
> refer to a field in the index and to another field that is not part of the
> specific local index. There is an open bug on it.
>
> BR,
>
> Yoav Sapir
>
>
> On 19 Sep 2017, at 18:21,
Hi Noam,
A few ideas:
1) Use a UNION instead of an OR and you may be able to use more than one
index for one query.
2) Include the columns you're referencing in the index to make it a covered
index [1].
3) Hint [2] the query to force the index to be used.
Thanks,
James
[1]
Hi Pradheep,
Would you be able to describe your use case and why you're salting? We
really only recommend salting if you have write hotspotting. Otherwise, it
increases the overall load on your cluster.
Thanks,
James
On Fri, Sep 8, 2017 at 9:13 AM, Pradheep Shanmugam <
write performance raised to 1k/s with the
>> tuning above.
>> I also tried using multi-thread for write, with 20 threads write
>> concurrently, the performance can raise to 5k/s, but won't raise any more
>> how ever threads increased.
>>
>> I didn't try wr
Hi Hef,
Have you had a chance to read our Tuning Guide [1] yet? There's a lot of
good, general guidance there. There are some optimizations for write
performance that depend on how you expect/allow your data and schema to
change:
1) Is your data write-once? Make sure to declare your table with the
Another option is to upgrade to 4.8 or later and use transactions (which
uses optimistic concurrency under the covers):
https://phoenix.apache.org/transactions.html
Thanks,
James
On Wed, Aug 30, 2017 at 9:53 AM, James Taylor <jamestay...@apache.org>
wrote:
> Upgrade to 4.9 or later an
Upgrade to 4.9 or later and use our atomic upsert command:
http://phoenix.apache.org/atomic_upsert.html
On Wed, Aug 30, 2017 at 9:38 AM Pradheep Shanmugam <
pradheep.shanmu...@infor.com> wrote:
> hi,
>
>
> we have a table in phoenix 4.4 which has a modifyrevision as a column..
>
> when a thread
The finest granularity we have available is to specify a guidepost width at
the table level with the GUIDEPOSTS_WIDTH=## property. This will take
effect for that table next time stats are collected (on a major compaction
or when collected manually through the UPDATE STATISTICS command).
On Tue,
Is the table salted? If the SALT_BUCKETS property was ever changed during
the life of the table, this could lead to duplicates. You could get the
same row key with a different salt byte. If not salted, are there any
global, mutable secondary indexes. A number of issues have been fixed which
will
Phoenix retrieves the server timestamp from the region server that hosts
the system catalog table and uses that as the timestamp of the puts when
you do an UPSERT VALUE (FYI, this behavior will change in 4.12 and we'll
use latest timestamp everywhere). I suspect the puts you're doing are going
to
Hi Pawan,
Why would you listen to someone about the future of Apache Phoenix who has
no involvement in or knowledge about the project? Michael Segel was banned
on the HBase lists for a while due to being too belligerent. He's not a
committer on any Apache project. Has he even contributed a patch?
The query for products is doing a point lookup based on the combinations
of product_id, sha_key, zip_code and then (I think) a skip scan is being
done for clicks based on the sha_key (based on the DYNAMIC SERVER FILTER).
You should be able to confirm this in a debugger through the scans that
take
I agree - the explain plan is confusing. Would be good to capture fixing
that in a JIRA. Would you mind sharing your DDL and queries too?
On Thu, Jul 27, 2017 at 3:55 PM, Josh Elser wrote:
>
>
> On 7/27/17 4:36 PM, Lew Jackman wrote:
>
>> I am joining two tables by using only
Only constant expressions are allowed for default value currently (as
mentioned by Juvenn). PHOENIX-3425 is for supporting expressions such as
CURRENT_DATE and NEXT VALUE FOR.
On Fri, Jul 14, 2017 at 7:22 AM Ankit Singhal
wrote:
> Phoenix 4.9 onwards you can specify
The Apache Phoenix team is pleased to announce the immediate availability
of the 4.11.0 release. Apache Phoenix enables SQL-based OLTP and
operational analytics for Apache Hadoop using Apache HBase as its backing
store and providing integration with other projects in the Apache ecosystem
such as
You shouldn't need to query an index table directly as Phoenix will use an
index automatically based on a determination by the optimizer that it's
more efficient to do so. If for diagnostic purposes you need to query the
index table directly (or any table that uses special characters in the
column
My recommendation: don't use salt buckets unless you have a monatomically
increasing row key, for example one that leads with the current date/time.
Otherwise you'll be putting more load (# of salt buckets more load worst
case) for bread-and-butter small-range-scan Phoenix queries.
Thanks,
James
Hi Michael,
Would it be possible for you to provide an example with DDL, query, and
explain plan?
Thanks,
James
On Fri, Jun 23, 2017 at 2:36 AM Mike Prendergast
wrote:
> We are currently using Phoenix 4.9 in EMR 5.6. We have been constructing
> queries that we would
Hi Pedro,
Before 4.10, it will be single client (though multi threaded). With 4.10
and above, the statement would run distributed across your cluster so
performance should improve. Note that if the source table is taking writes
while the UPSERT SELECT is running, the statement would miss those
point me to example or to related
> source code of Phoenix API?
>
> On 16 Jun 2017, at 03:41, James Taylor <jamestay...@apache.org> wrote:
>
> No, not unless you write the code that keeps the index in sync yourself,
> attach the metadata needed by index maintenance coproces
No, not unless you write the code that keeps the index in sync yourself,
attach the metadata needed by index maintenance coprocessor in your update
code, or use the Phoenix update APIs which do this for you.
On Thu, Jun 15, 2017 at 5:29 PM, Batyrshin Alexander <0x62...@gmail.com>
wrote:
> We
Hi Alexander,
We support indexes on views which is essentially what you're asking for.
Thanks,
James
On Thu, Jun 8, 2017 at 1:28 AM Batyrshin Alexander <0x62...@gmail.com>
wrote:
> Hello,
> Is there any plans to implement partial indexes (CREATE INDEX WHERE
> predicate)?
be either "index" or "value" (what to delete by)
> and the last would be the index or value to delete
>
> Regards,
> Cheyenne
>
> On Tue, Jun 6, 2017 at 6:17 PM, James Taylor <jamestay...@apache.org>
> wrote:
>
>> Please feel free to file
We don't have APIs for what you're trying to do. Our APIs are JDBC. For
these low level kinds of things, I'd recommend reading and understanding
the source code.
Thanks,
James
On Mon, Jun 5, 2017 at 1:09 PM Cheyenne Forbes <
cheyenne.osanu.for...@gmail.com> wrote:
> Can anyone please help?
>
>
Thanks, Ash. We welcome any doc enhancements. See
http://phoenix.apache.org/building_website.html on how to do this.
James
On Wed, May 31, 2017 at 7:54 AM, Ash N <742...@gmail.com> wrote:
> Hi Samarth,
>
> Your comment is quite decisive and actionable.
>
> I'd like to edit the documentation
Hello,
The index rebuild process is incremental, based on the time stamp at which
the index write failure originally occurred.
Thanks,
James
On Wed, May 31, 2017 at 2:22 AM 金砖 wrote:
> hi, everyone:
>
> I read from document there are 3 levels of consistency.
>
>
Answered over on SO
On Tue, May 23, 2017 at 3:34 PM, Barrett Strausser
wrote:
> Crossposted to SO
>
> https://stackoverflow.com/questions/44144925/apache-
> phoenix-current-time-gives-npe
>
FWIW, we're exposing a way to do snapshot reads (PHOENIX-3744), starting
with our MR integration (on top of which the Spark integration is built)
for our 4.11 release. This is about as close as you can get to reading HDFS
directly while still taking into account non flushed HBase data.
Thanks,
But hopefully Hortonworks hears (from you and many others) that you'd like
to see HDP move to a newer version of Phoenix. The 4.7 release is four
releases back from the current 4.10 release.
Any ballpark timeframe for an HDP release with the latest Phoenix release?
On Fri, May 19, 2017 at 7:56
t; An d keep the offsets for each page in the server side, for a while...
> (the client might not need all of them, might also never ask for next page)
>
> http://www.oracle.com/technetwork/java/valuelisthandler-142464.html
>
> On May 18, 2017 20:02, "James Taylor" <james
Yes, it's expected that query performance would degrade as the offset
increases. The best Phoenix can do for OFFSET is to scan the rows and count
them until the offset is reached. Use the row value constructor technique
instead to prevent this: https://phoenix.apache.org/paged.html
On Thu, May
Hi Ash,
Typically users come up with a composite primary key based on event data
that is unique. For example: . If you
don't have an event ID from an external source, perhaps you can either use
a cyclical sequence or a random number (using our RAND built-in function).
Would that give you the
Did you commit the data? I'd recommend turning auto commit on for the
connection before running the delete statement.
Thanks,
James
On Fri, Apr 28, 2017 at 12:36 PM, Michael Young wrote:
> This is probably a newbie question, but has anyone run across a case where
> a
;
>>>>> I am creating a plugin that uses Lucene to index text fields and I
>>>>> need to access *getConf()* and *getFilesystem()* of *HRegion, *the
>>>>> Lucene indexes are split with the regions so I need " *HRegion
>>>>> MyVar; ", *I am po
Shorter answer is "no". Your UDF may be executed on the client side as well
(depending on the query) and there is of course no HRegion available from
the client.
On Tue, Apr 18, 2017 at 11:10 AM Sergey Soldatov
wrote:
> Well, theoretically there is a way of having a
No, that's not possible.
Thanks,
James
On Fri, Apr 14, 2017 at 9:48 AM Cheyenne Forbes <
cheyenne.osanu.for...@gmail.com> wrote:
> If the UDF will be used only in select queries "SELECT my_udf(column) FROM
> table WHERE id = 123" would it be possible to access the MyCoprocessor
> instance on the
ing and what might cause
> performance issues or errors during usage.
>
> Also, any other way to analyze these logs besides Splunk?
>
> On Tue, Apr 11, 2017 at 4:05 PM, James Taylor <jamestay...@apache.org>
> wrote:
>
>> FWIW, we use our metric framework[1] to collect t
The non standard @ syntax is not supported by Phoenix: (@rn:= if((@item =
t.item_id) AND (@type = t.type),@rn + 1,
if((@item:=t.item_id) AND (@type:=t.type),1,1)
Also, I'd recommend breaking that query up into multiple queries to
pinpoint any other non supported constructs.
On
FWIW, we use our metric framework[1] to collect that information,
outputting one log line per statement or query and then Splunk to look at
it.
[1] http://phoenix.apache.org/metrics.html
On Tue, Apr 11, 2017 at 3:24 PM Michael Young wrote:
> Yes, the tracing webapp is
gt;> Releasing is a different matter. I'm not sure how Phoenix releases are
> >> currently handled, but for a Python library, it makes sense to upload
> >> the package to PyPI since that's from where most Python developers are
> >> getting packages from. Who w
What other queries do you want to be fast? What version of Phoenix are you
using? Have you seen our new Tuning Guide [1]?
You could try moving OBJ_ID to end of PK and adding OBJ_GRP as the leading
PK column (it's fine to include it in PK even if it's optional).
Your index should be used, though.
Hi Yunus,
No, unfortunately that won't work between minor releases. The server side
needs to be upgraded (first) followed by the client side. See here [1] for
more info.
Thanks,
James
[1] http://phoenix.apache.org/upgrading.html
On Tue, Apr 4, 2017 at 9:01 AM wrote:
No.
On Tue, Mar 28, 2017 at 10:19 AM Cheyenne Forbes <
cheyenne.osanu.for...@gmail.com> wrote:
> So I cant get neither the name of the table, the name of the column, the
> id of the row nor the region that "ptr" belongs to after calling:
>
>
>
>
>
> *Expression arg =
No, that's not possible. Keep in mind that these expressions may be
evaluated on the client or the server (it depends on the query). The could
be evaluated on post aggregated rows too where the idea of which region no
longer makes sense.
On Sat, Mar 25, 2017 at 10:15 PM Cheyenne Forbes <
Phoenix supports JDBC, not any kind of scripting language like PLSQL or
other variant. You can loook up online lots of examples of using
PreparedStatement to bind variables in a query.
On Tue, Mar 28, 2017 at 5:34 AM Cheyenne Forbes <
cheyenne.osanu.for...@gmail.com> wrote:
> Can someone show an
racle API . There is getIntArray method to cast
> Oracle Array to int[] .Phoenix array does not hava one
>
> 在 2017年3月27日,22:38,James Taylor <jamestay...@apache.org> 写道:
>
> The JDBC array is not an ArrayList. See
> http://download.oracle.com/otn_hosted_doc/jdeveloper/905/jdbc-javadoc/oracle
The JDBC array is not an ArrayList. See
http://download.oracle.com/otn_hosted_doc/jdeveloper/905/jdbc-javadoc/oracle/sql/ARRAY.html
On Mon, Mar 27, 2017 at 3:50 AM 袁康(基础平台部)
wrote:
> When I query phoenix in java project
>
> occur:java.lang.ClassCastException:
>
Bernard - the row value constructor syntax will work in the version you're
using and it's a more scalable way of doing paging over HBase data.
Thanks,
James
On Fri, Mar 24, 2017 at 3:16 AM, rajeshb...@apache.org <
chrajeshbab...@gmail.com> wrote:
> Offset is supported from Phoenix 4.8.0+
t; > part it worked. Definitely feel free to report any issues you run into:
> > https://bitbucket.org/lalinsky/python-phoenixdb. It would be nice to
> make
> > this a "first-class citizen".
> >
> > James Taylor wrote:
> >
> >> Thanks, Dimitri. I c
The Apache Phoenix team is pleased to announce the immediate availability
of the 4.10.0 release. Apache Phoenix enables SQL-based OLTP and
operational analytics for Hadoop using Apache HBase as its backing store
and providing integration with other projects in the ecosystem such as
Spark, Hive,
The bug is HBASE-17587.
Thanks,
James
On Tue, Mar 21, 2017 at 7:54 AM Mike Prendergast
wrote:
> According to the link below, it seems that Phoenix 4.9 won't work with
> Hbase 1.3.0, and they are waiting for 1.3.1 for a bug fix. What is the
> specific bug mentioned
UNNEST is only supported in our calcite branch currently. There's been no
release yet with support for it. If you're interested, I'd encourage you to
get involved in the Phoenix/Calcite integration work or implement support
for it in our 4.x branches.
Thanks,
James
On Sun, Mar 19, 2017 at 4:59
The DATE and TIME types in Phoenix are more or less the same. They store
their data as an 8 byte long Epoch time and maintain millisecond
granularity.
The TIMESTAMP stores nanosecond data in a 12 bytes - an 8 byte long Epoch
time plus 4 bytes for nanoseconds. Unless you need nanosecond
I'm excited to announce that the 2nd Annual Apache Phoenix conference,
PhoenixCon 2017 will take place the day after HBaseCon in San Francisco on
Tuesday, June 13th from 10:30am-6pm. For more details, including to RSVP
and submit a talk proposal, click here:
Hi Marcin,
There's a few things going on here:
1) An index created on a view over HBase data won't be maintained by
Phoenix. You'd need to maintain it yourself through some other external
means. If you create a table that maps to your HBase data, then it will be
maintained.
2) An index would only
Sounds like you're using the wrong jar on the server side. Or perhaps
you're using a vendor distribution (in which case you'll need to use their
supported Phoenix version). Please review the installation instructions as
everything you need is in the one single jar.
Thanks,
James
On Wed, Feb 1,
e link provided. You linked the gmail sent mail
> address i guess. Gmail can't redirect me to that and logs me back to my
> inbox. Could you please repost it?
>
> Thanks.
>
> Regards.
>
> On Wed, Jan 25, 2017 at 9:44 PM, James Taylor <jamestay...@apache.org>
> wrote:
Please see this thread[1] for a description of how to do time series in
Apache Phoenix.
[1] https://mail.google.com/mail/u/0/#sent/15985640ed6c5e86
On Wed, Jan 25, 2017 at 4:26 AM, Dhawal Arora wrote:
> Hi,
>
> I'm trying to map a clickstream data into Hbase Schema
You can set the global phoenix.query.timeoutMs config in your client side
hbase-site.xml [1], or set the same in the Connection Properties for it to
apply just for that connection, or you can set it per query with the
Statement.setQueryTimeout method:
Yes, the consequences could be bad as you're bypassing the API that drops a
table (which does more than just update the SYSTEM. CATALOG). Please see
the recommendations in the previous emails.
On Mon, Jan 16, 2017 at 10:45 PM Batyrshin Alexander <0x62...@gmail.com>
wrote:
> I've discovered that
Couple of other options:
1) connect with a CURRENT_SCN [1] earlier than any timestamp for your data.
For example, if you know you created the table a month ago, then use an SCN
value of the epoch time for 1/1/2016. Then issue the DROP TABLE command
from that connection and no data will be removed.
Hi Udit,
You'll need to use Phoenix APIs when you update your data if you want
Phoenix to keep you index in sync with your table.
Thanks,
James
On Tue, Jan 3, 2017 at 1:32 PM, Udit Mehta wrote:
> Hi All,
>
> I am facing an issue while working with secondary indexes on a
This definitely sounds like a serious problem. Agreed, it should be closed
immediately. Also, a minor compaction shouldn't trigger any stats
collection logic. Please file a JIRA and we'll get this fixed in the next
release. As always, patches are welcome.
Thanks,
James
On Wed, Dec 28, 2016 at
Hi Jeremy,
Please file a JIRA for VIEW_INDEX_ID to be an int. The code would just need
to be tolerant when reading the data if the length is two byte short versus
four byte int. At write time, we'd just always write an int. If you could
put together a patch that maintains backward compatibility,
Yes - use :1 instead of ?1
On Thu, Dec 8, 2016 at 12:11 PM Cheyenne Forbes <
cheyenne.osanu.for...@gmail.com> wrote:
> So instead of doing:
>
> query("select from table where c1 = ? or c2 = ?", [my_id, my_id])
>
> i would do:
>
> query("select from table where c1 = ?1 or c2 = ?1",
Thanks for narrowing down the cause, Karthick. Please file a JIRA with this
information.
James
On Wed, Dec 7, 2016 at 12:45 PM, KARTHICK Duraisamy Soundararaj <
ksoundara...@groupon.com> wrote:
> I am able to create a secondary index on case-insensitive schema.table and
> everything works
Hi Venkat,
Did you ever update the SYSTEM.CATALOG table manually? If so, did you set
the CURRENT_SCN property on the connection so that the timestamp of the
modified rows continue to have the required timestamp
(MetaDataProtocol.MIN_SYSTEM_TABLE_TIMESTAMP_4_6_0 or 9 in this case)? It
sounds like
Looks like a bug to me. If you can reproduce the issue outside of Python
phoenixdb, using the Query Server directly, please file a JIRA. If not, it
sounds like a Phython phoenixdb issue.
Thanks,
James
On Tue, Dec 6, 2016 at 8:58 AM, Mark Heppner wrote:
> I encountered
Have you tried local indexes?
On Mon, Dec 5, 2016 at 2:35 PM Neelesh wrote:
> Hello,
> When a region server is under stress (hotspotting, or large replication,
> call queue sizes hitting the limit, other processes competing with HBase
> etc), we experience latency spikes
Apache Phoenix enables OLTP and operational analytics for Apache Hadoop
through SQL support using Apache HBase as its backing store and providing
integration with other projects in the ecosystem such as Apache Spark,
Apache Hive, Apache Pig, Apache Flume, and Apache MapReduce.
We're pleased to
workin fine when using "=" and
> it is not working with LIKE operator.
>
> It seems like parameters passed to LIKE operator are treated in a bit
> different way.
>
>
>
> I am using DBeaver to test my queries and when I am using this notation
> : (:1) or wirh paramet
Not sure of the JIRA, but it sounds familiar. Try searching for it here:
https://issues.apache.org/jira/browse/PHOENIX
On Sat, Nov 26, 2016 at 1:22 PM Neelesh <neele...@gmail.com> wrote:
> Thanks James! Is there a jira ref for the fix?
>
> On Nov 26, 2016 11:50 AM, "Jam
in the connection or explicitly make the commit?
>
> On Sat, Nov 26, 2016 at 3:13 PM, James Taylor <jamestay...@apache.org>
> wrote:
>
> Are you committing the data after the upsert (either by turning on auto
> commit on the connection or doing an explicit commit)?
>
>
Are you committing the data after the upsert (either by turning on auto
commit on the connection or doing an explicit commit)?
On Fri, Nov 25, 2016 at 12:20 PM Matt L wrote:
> Hello All!
>
> Through Zeppelin I am trying to insert data into a table created by
> Phoenix:
>
>
Use : instead, like this :1
Thanks,
James
On Fri, Nov 18, 2016 at 5:28 AM Bartłomiej Niemienionek <
b.niemienio...@gmail.com> wrote:
> Hi,
>
>
>
> I am trying to use Phoenix and I am facing some problems with LIKE
> operator when used in the prepared statement with bind parameters.
>
> I don’t
gt; It works with no column family, but I expect that I do not need to make
> sure column names are unique across different column families.
>
>
>
> Xindian
>
>
>
>
>
> *From:* James Taylor [mailto:jamestay...@apache.org
> <jamestay...@apache.org>]
> *Sent:*
Have you tried without the column family name? Unless the column names are
not unique across all column families, you don't need to include the column
family name.
Thanks,
James
On Tue, Nov 8, 2016 at 2:19 PM, Long, Xindian
wrote:
> I have a table with multiple column
It sounds like PHOENIX-3178, but if not, please file a new JIRA with a way
to reproduce it. Patches are most welcome too, of course.
Thanks,
James
On Thu, Nov 3, 2016 at 4:38 PM, Julian Jaffe
wrote:
> When I execute an upsert with filters through the Phoenix JDBC
Awesome, Lukas! Thanks so much - this is great!
On Thu, Nov 3, 2016 at 2:14 PM, Lukáš Lalinský wrote:
> Hello,
>
> I have released new version of the Python client library for working with
> Phoenix. Thanks to Mark Heppner it now uses the Protocol Buffers interface
> instead
CdOABUrV8Pw
>> <https://www.linkedin.com/profile/view?id=AAEWh2gBxianrbJd6zP6AcPCCdOABUrV8Pw>*
>>
>>
>>
>> http://talebzadehmich.wordpress.com
>>
>>
>> *Disclaimer:* Use it at your own risk. Any and all responsibility for
>> any
Keep in mind that the CsvBulkLoadTool does not handle updating data
in-place. It's expected that the data is unique by row and not updating
existing data. If your data is write-once/append-only data, then you'll be
ok, but otherwise you should stick with using the JDBC APIs.
You're free to just
That's awesome, Cheyenne! Is it packaged as an ODBC client or is the API
different? If you send us a link to your Github repo and docs, we'll get
links up on the website. Would be great to get an ODBC driver contributed
to Phoenix and included in our distro.
Thanks,
James
On Sunday, October 23,
Hi Mich,
Phoenix indexes are only updated if you use Phoenix APIs to input the data.
Thanks,
James
On Saturday, October 22, 2016, Ted Yu wrote:
> The first statement creates index, not view.
>
> Can you check ?
>
> Cheers
>
> > On Oct 22, 2016, at 1:51 AM, Mich Talebzadeh
Hi Sanooj,
I'm not sure what you mean by "loading data in our HBase table into
in-memory", but Phoenix queries tables in parallel, even within a region
depending on how you've configured statistics and guideposts as described
here: http://phoenix.apache.org/update_statistics.html
Thanks,
James
FYI, a couple of timestamp related features that Phoenix supports today
include;
- specify/filter on timestamp of a row:
http://phoenix.apache.org/rowtimestamp.html
- query as of a past timestamp:
On Fri, Oct 14, 2016 at 12:37 AM, Antonio Murgia
wrote:
> We tried with an Upsert from select, but we ran into some memory issue
> from the phoenix side.
>
> Do you have any suggestion to perform something like that?
>
You can try setting auto commit to true on the
Hi Zack,
Phoenix doesn't support an ALTER TABLE statement that allows this, but you
can modify the SYSTEM.CATALOG directly with an UPSERT VALUES call if you're
careful. Make sure to:
- Create a snapshot of the SYSTEM.CATALOG just in case something goes wrong
- Ensure that other modifications
Hi Mich,
I'd encourage you to use this mechanism mentioned by Josh:
Another option is to use Phoenix-JDBC from within Spark Streaming. I've got
a toy example of using Spark streaming with Phoenix DataFrames, but it
could just as easily be a batched JDBC upsert.
Trying to write directly to HBase
Hi Sanjooj,
What version of Phoenix? Would you mind filing a JIRA with steps to
reproduce the issue?
Thanks,
James
On Friday, October 7, 2016, Sanooj Padmakumar wrote:
> Hi All
>
> We get mutation state related error when we try altering a table to which
> views are added.
FWIW, you can use an "alter view drop column" statement after you've
created your view to remove columns, though.
Thanks,
James
On Friday, October 7, 2016, Mich Talebzadeh
wrote:
> thanks Ankit
>
> Dr Mich Talebzadeh
>
>
>
> LinkedIn *
>
ease let me know if there is any other better approach that i am missing.
>
> On Sun, Oct 2, 2016 at 3:50 PM, James Taylor <jamestay...@apache.org
> <javascript:_e(%7B%7D,'cvml','jamestay...@apache.org');>> wrote:
>
>> Hi Anil,
>> Make sure you're not sharing th
Hi Anil,
Make sure you're not sharing the same Connection between multiple threads
as it's not thread safe.
Thanks,
James
On Sunday, October 2, 2016, anil gupta wrote:
> Hi,
>
> We are running HDP2.3.4(HBase 1.1 and Phoenix 4.4). I have a MapReduce job
> thats writing
101 - 200 of 667 matches
Mail list logo