[
https://issues.apache.org/jira/browse/PHOENIX-1071?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14389417#comment-14389417
]
ASF GitHub Bot commented on PHOENIX-1071:
-----------------------------------------
Github user apurtell commented on the pull request:
https://github.com/apache/phoenix/pull/59#issuecomment-88248990
I'm not a Spark expert @JamesRTaylor . I skimmed the latest. Allowing
builds with JDK 1.7 would have been the big change I'd have recommended, and
it's already been done.
I checked out this PR and ran a build, which completed. I was able to run
the unit tests of the new module from the Maven command line, on Linux FWIW:
$ mvn -DskipTests clean install
$ mvn test -rf :phoenix-spark
[...]
- Can create valid SQL
- Can convert Phoenix schema
- Can create schema RDD and execute query
- Can create schema RDD and execute query on case sensitive table (no
config)
- Can create schema RDD and execute constrained query
- Using a predicate referring to a non-existent column should fail
- Can create schema RDD with predicate that will never match
- Can create schema RDD with complex predicate
- Can query an array table
- Can read a table as an RDD
- Can save to phoenix table
- Can save Java and Joda dates to Phoenix (no config)
- Not specifying a zkUrl or a config quorum URL should fail
Run completed in 1 minute, 12 seconds.
Total number of tests run: 13
Suites: completed 2, aborted 0
Tests: succeeded 13, failed 0, canceled 0, ignored 0, pending 0
All tests passed.
With 7u75 I run out of PermGen running PhoenixRDDTest, but fixed that:
diff --git a/phoenix-spark/pom.xml b/phoenix-spark/pom.xml
index 5c0c754..21baa16 100644
--- a/phoenix-spark/pom.xml
+++ b/phoenix-spark/pom.xml
@@ -503,6 +503,7 @@
<configuration>
<parallel>true</parallel>
<tagsToExclude>Integration-Test</tagsToExclude>
+ <argLine>-Xmx3g -XX:MaxPermSize=512m
-XX:ReservedCodeCacheSize=51
</configuration>
</execution>
<execution>
The unit tests are not robust against parallel execution with other HBase
or Phoenix test suite invocations on the same host, but this can be fixed with
a followup issue with random ports and rebinding
LGTM for a commit to trunk with some minor follow-ups.
> Extend the org.apache.spark.sql.sources.RelationProvider and have
PhoenixDatasource.
Maybe we should split this work up. The integration as-is is directly
useful on its own. The SparkSQL integration nice-to-have can be additional work
on a new JIRA / PR?
> Provide integration for exposing Phoenix tables as Spark RDDs
> -------------------------------------------------------------
>
> Key: PHOENIX-1071
> URL: https://issues.apache.org/jira/browse/PHOENIX-1071
> Project: Phoenix
> Issue Type: New Feature
> Reporter: Andrew Purtell
>
> A core concept of Apache Spark is the resilient distributed dataset (RDD), a
> "fault-tolerant collection of elements that can be operated on in parallel".
> One can create a RDDs referencing a dataset in any external storage system
> offering a Hadoop InputFormat, like PhoenixInputFormat and
> PhoenixOutputFormat. There could be opportunities for additional interesting
> and deep integration.
> Add the ability to save RDDs back to Phoenix with a {{saveAsPhoenixTable}}
> action, implicitly creating necessary schema on demand.
> Add support for {{filter}} transformations that push predicates to the server.
> Add a new {{select}} transformation supporting a LINQ-like DSL, for example:
> {code}
> // Count the number of different coffee varieties offered by each
> // supplier from Guatemala
> phoenixTable("coffees")
> .select(c =>
> where(c.origin == "GT"))
> .countByKey()
> .foreach(r => println(r._1 + "=" + r._2))
> {code}
> Support conversions between Scala and Java types and Phoenix table data.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
