Preetpal Singh created PHOENIX-1497:
---------------------------------------
Summary: HIVE UDTF (OR Cassandra Clustering column) like
functionality in Phoenix
Key: PHOENIX-1497
URL: https://issues.apache.org/jira/browse/PHOENIX-1497
Project: Phoenix
Issue Type: New Feature
Reporter: Preetpal Singh
Ability to access multiple column qualifiers (CQs) within a HBASE rowkey and
its associate values (CV) as two columns in Phoenix.
This kind of functionality exists in CQL 3.0 (Cassandra) by way of a clustering
column. HIVE provides a function called Explode or ability to write any custom
UDTF for purpose of transposing CQs and associated values into a set of
relational columns.
HIVE UDTF --
https://cwiki.apache.org/confluence/display/Hive/LanguageManual+UDF#LanguageManualUDF-Built-inTable-GeneratingFunctions(UDTF)
Cassandra Clustering column --
http://www.datastax.com/documentation/cql/3.0/cql/cql_reference/create_table_r.html
(read clustering column part)
Elaborate description
1 For a given rowkey R1, which has associated to it CQs (Disregard column
family concept for a moment), following would be the data in HBASE.
c1:v1, c2:v2, c3:v3 so and so forth (suppose 1000 of these).
2. In Phoenix we should be able to to run a query like this
select rowkey, cola, colb from tablea where cola > c5
This should present as output the following
Rowkey COLA COLB
R1 C1 V1
R1 C2 V2
R1 C3 V3
R1 C4 V4
R1 C5 V5
R1 C6 V6
R1 C7 V7
R1 C8 V8
R2 C9 V9
R2 C10 V10
R2 C11 V11
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
