phoenix-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bruno Dumon <bruno.du...@gmail.com>
Subject PK-based random access performance issue
Date Sun, 02 Feb 2014 07:54:53 GMT
Hi,

I have a simpe table like this:

create table CUSTOMER (
  ID BIGINT NOT NULL PRIMARY KEY,
  DATA.FIRST_NAME VARCHAR,
  DATA.LAST_NAME VARCHAR,
  DATA.GENDER VARCHAR,
  DATA.CITY VARCHAR
)
SALT_BUCKETS = 20

Doing simple PK-based queries on this table, like this:

select * from customer_dna where ID = 1

takes an extermely long time, while it is basically the equivalent of a Get
on HBase.

Explain says:

CLIENT PARALLEL 56-WAY SKIP SCAN ON 20 KEYS OVER CUSTOMER [0,1] - [19,1]

which seems to suggest it does full table scans.

I tried with 2.2.2 and 2.2.3.

Any ideas?

Bruno.

Mime
View raw message