Hi James,
Thanks for your reply, but further debugging pointed out that it was
something I was doing wrong on the client side and was able to resolve it.
Sorry for the inconvenience.
Cheers,
Anirudha
On Wed, Jul 1, 2015 at 9:45 PM, James Taylor wrote:
> Hi Anirudha,
> No, there's no difference
Yufan,
What version of Phoenix are you using?
Thanks,
James
On Wed, Jul 1, 2015 at 2:34 PM, Yufan Liu wrote:
> When I made more tests, I find that this problem happens after table got
> split.
>
> Here is the DDL I use to create table and index:
> CREATE TABLE IF NOT EXISTS t1 (
> uid BIGINT NOT
Hi Nishant,
Refer to HBase wiki for multiple column families. As per my experience, don't
try to have more than 2-3 column family. Also group the column in column
families on basis of access pattern.
If you don't have an access where you can avoid reading a column family then
you would not ga
Thanks Puneet and James for your responses.
Date is not recommended as first part of rowkey. It will create issue
during write operation. In real production scenario we will have more data
and will have more values for column1 and column2.
Will try other things today. Lets see how much I can achi
When I made more tests, I find that this problem happens after table got
split.
Here is the DDL I use to create table and index:
CREATE TABLE IF NOT EXISTS t1 (
uid BIGINT NOT NULL,
timestamp BIGINT NOT NULL,
eventName VARCHAR
CONSTRAINT my_pk PRIMARY KEY (uid, timestamp)) COMPRESSION='SNAPPY';
PHOENIX-2094 created.
Covering the queries isn’t a viable option right now, the table is 5TB already
and we have multiple indexes where we are trying to optimize case-insensitive
queries. Fortunately we can brute-force the plan with the JOIN & subselect to
get what we need near-term.
Bryan G.
If you could put a complete test (including your DDL and upsert of data),
that would be much appreciated.
Thanks,
James
On Wed, Jul 1, 2015 at 11:20 AM, Yufan Liu wrote:
> I have tried to use query "SELECT timestamp FROM t1 ORDER BY timestamp
> DESC NULLS LAST LIMIT 1". But it still returns the
I have tried to use query "SELECT timestamp FROM t1 ORDER BY timestamp DESC
NULLS LAST LIMIT 1". But it still returns the same unexpected result. There
seems to be some internal problems related.
2015-06-30 18:03 GMT-07:00 James Taylor :
> Yes, reverse scan will be leveraged when possible. Make y
Phoenix doesn’t support automatically using index on a non-covered query. Check
out https://phoenix.apache.org/secondary_indexing.html for some examples. Your
query may have been faster due to more data being cached.
The 2-step plan you’re looking for in this case looks more like this one where
Baahu,
We're having a difficult time reproducing the StackOverflowError you
encountered over on PHOENIX-2074. Do you think you could help us reproduce
it? Maybe you can upload a test case and/or a CSV file with some sample
data that reproduces it?
Thanks,
James
On Tue, Jun 23, 2015 at 12:24 AM, Ba
Also, try separating your columns into multiple column families to prevent
having to scan past your 75+ column qualifiers for every query.
On Wed, Jul 1, 2015 at 4:47 AM, Puneet Kumar Ojha wrote:
> Yes …Salting will improve the scan performance. Try with numbers 5,10,20
> . As I do not know abo
Hi Anirudha,
No, there's no difference between using sqlline and a Java client (sqlline
is a Java client).
If you can put together a test with your original OR query and some sample
data that exhibits the behavior, please let us know.
Thanks,
James
On Wed, Jul 1, 2015 at 3:40 AM, Anirudha Khanna
I seems
- CREATE TABLE returns the error "Table already exists" if you earlier
created and dropped a view for the HBase table (bug?).
- but one can actually run CREATE TABLE for an existing hbase table
according to the documentation
I can't be sure for the moment as the CREATE TABLE statement on a
Hi,
I got an existing HBase table, so I mapped to phoenix using a *view*. I can
select, create index, so I am happy.
Now I want to add a row (I assume it is not compulsory to have all column
values defined - would be boring otherwise with numerous columns):
upsert into "table" ("family1"."column
Hi,
I want to perform :
SELECT * FROM "table" where "family"."column1" = 'value'
Running an EXPLAIN on this request before creating an index on a column
gives :
CLIENT PARALLEL 1-WAY FULL SCAN OVER table
SERVER FILTER BY family.column1 = 'value'
Looks OK.
Then I simply :
CREATE INDEX "table_I
After some investigation, I think this is a permissions issue.
If I run as ‘hbase’, this works consistently.
FYI
From: Riesland, Zack
Sent: Wednesday, July 01, 2015 7:25 AM
To: user@phoenix.apache.org
Subject: Help interpretting CsvBulkLoader issues?
After using the CsvBulkLoader successfully f
Yes …Salting will improve the scan performance. Try with numbers 5,10,20 . As I
do not know about the cluster details.
Increase scanner caching to 10.
Check if SNAPPY is working …I hope you need to put the jars classpath as well.
Since the cardinality of the col1 and col2 fields is very sma
After using the CsvBulkLoader successfully for a few days, I’m getting some
strange behavior this morning.
I ran the job on a fairly small ingest of data (around 1/2 billion rows).
It seemed to complete successfully. I see this in the logs:
Phoenix MapReduce Import
Upser
HI Puneet/Martin,
Thanks for your response. Please see my answer as below.
I have not specified any salt bucket. I have created Phoenix View on
existing Hbase Table. Can I specify Salt bucket for Phoenix View?
After loading Hbase data I alter table to use SNAPPY Compression. Are you
talking abou
Hi,
I have a simple query like,
*SELECT tb1."id" AS "id"*
*FROM "ACCOUNTS" tb1*
*WHERE tb1."id" = 87*
* OR tb1."id" = 89 LIMIT 1*
which when running through JAVA using jdbc driver returns me only one row.
Whereas there are two rows in the table with those ids.
The explain plan of this (fro
It sounds like you are scanning rather than getting rows based on a known
row id. Am I wrong?
One thing I am currently trying is to have indexed columns and "hot"
content in one column family and let "cold" content in another family. It
speed up scanning the table when you need to
Le mer. 1 juil.
· Need to know the salt buckets you have created.
· Compression Used or Not.
· Number Set for returning result ..it should be increased.
· Keep deleted cells should be false. If you delete records then it
slows the response time.
Please provide the query which y
22 matches
Mail list logo
