Apache Zeppelin - Set default interpreter

问题

In Zeppelin, at each row I am having to provide the interpreter at each row. Is there a way to set the interpreter for the whole session.

%pyspark 

import re

Took 0 seconds.

import pandas as pd

console :1: error: '.' expected but identifier found. import pandas as pd

%pyspark

import pandas as pd

Took 0 seconds.

How do I set the interpreter for the whole session?

回答1:

The Spark Interpreter group currently has 4 interpreter as listed here...

https://zeppelin.incubator.apache.org/docs/0.5.0-incubating/interpreter/spark.html

The default interpreter is %spark and default interpreter is selected based on the order of the interpreter listed in the zeppelin.interpreters property in zeppelin-site.xml config file.

The current order of interpreter in your zeppelin-site.xml (zeppelin.interpreters property) will be this ...

org.apache.zeppelin.spark.SparkInterpreter,org.apache.zeppelin.spark.PySparkInterpreter

Modify this to ...

org.apache.zeppelin.spark.PySparkInterpreter, org.apache.zeppelin.spark.SparkInterpreter

and restart Zeppelin (zeppelin-daemon.sh restart)

This will make %pyspark as default interpreter.

Thanks

回答2:

The above answer did not work on recent Zeppelin versions.

To set the default interpreter, check /etc/zeppelin/conf/interpreter.json and look for something like:

...
{
  "name": "spark",
  "class": "org.apache.zeppelin.spark.SparkInterpreter",
  "defaultInterpreter": true,
  "editor": {
    "language": "scala",
    "editOnDblClick": false
  }
},
...
{
  "name": "pyspark",
  "class": "org.apache.zeppelin.spark.PySparkInterpreter",
  "defaultInterpreter": false,
  "editor": {
    "language": "python",
    "editOnDblClick": false
  }
}

Now everything seems trivial. We just need to change the defaultInterpreter of spark to false, and defaultInterpreter of pyspark to true.

And then restart the zeppelin (sudo stop zeppelin; sudo start zeppelin).

回答3:

Even Fishball's answer for recent Zeppelin seems outdated. My conf/interpreter.json came with spark as default ("defaultInterpreter": true) and python/pyspark as not ("defaultInterpreter": false) and yet Zeppelin picked up python/pyspark as default. And in my case, I wanted spark over pyspark.

The solution was to just drag and drop interpreters in Zeppelin web console's interpreter binding section.

来源：https://stackoverflow.com/questions/33834401/apache-zeppelin-set-default-interpreter

标签

apache-zeppelin