[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki


[ Back to the navigation ]

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
spark:using-python [2015/10/23 14:24]
ufal
spark:using-python [2022/12/14 13:25] (current)
straka [Usage Examples]
Line 8: Line 8:
 <file>pyspark</file> <file>pyspark</file>
  
-Better interactive shell with code completion using ''ipython'' (installed everywhere on cluster; ask our IT if you want to have it installed on your workstations too) can be started using: +Better interactive shell with code completion using ''ipython3'' (do `pip3 install --user ipython` if you do not have it) can be started using: 
-<file>IPYTHON=pyspark</file>+<file>PYSPARK_DRIVER_PYTHON=ipython3 pyspark</file>
  
 As described in [[running-spark-on-single-machine-or-on-cluster|Running Spark on Single Machine or on Cluster]], environmental variable ''MASTER'' specifies which Spark master to use (or whether to start a local one). As described in [[running-spark-on-single-machine-or-on-cluster|Running Spark on Single Machine or on Cluster]], environmental variable ''MASTER'' specifies which Spark master to use (or whether to start a local one).
Line 19: Line 19:
    .flatMap(lambda line: line.split())    .flatMap(lambda line: line.split())
    .map(lambda word: (word, 1))    .map(lambda word: (word, 1))
-   .reduceByKey(lambda c1,c2: c1+c2) +   .reduceByKey(lambda c1, c2: c1+c2) 
-   .sortBy(lambda (word,count)count, ascending=False)+   .sortBy(lambda word_countword_count[1], ascending=False)
    .take(10))    .take(10))
 </file> </file>
  
-  * run interactive shell using existing Spark cluster (i.e., inside ''spark-qrsh''), or start local Spark cluster using as many threads as there are cores if there is none: +  * run interactive shell using existing Spark cluster (i.e., inside ''spark-srun''), or start local Spark cluster using as many threads as there are cores if there is none: 
-  <file>IPYTHON=pyspark</file>+  <file>PYSPARK_DRIVER_PYTHON=ipython3 pyspark</file>
   * run interactive shell with local Spark cluster using one thread:   * run interactive shell with local Spark cluster using one thread:
-  <file>MASTER=local IPYTHON=pyspark</file> +  <file>MASTER=local PYSPARK_DRIVER_PYTHON=ipython3 pyspark</file> 
-  * start Spark cluster (10 machines, 1GB RAM each) on SGE and run interactive shell: +  * start Spark cluster (10 machines, 2GB RAM each) on Slurm and run interactive shell: 
-  <file>IPYTHON=spark-qrsh 10 1G pyspark</file>+  <file>PYSPARK_DRIVER_PYTHON=ipython3 spark-srun 10 2G pyspark</file>
  
-Note that ''IPYTHON'' variable can be left out or specified in ''.bashrc'' (or similar).+Note that ''PYSPARK_DRIVER_PYTHON'' variable can be left out or specified in ''.bashrc'' (or other configuration files).
  
  
Line 59: Line 59:
    .flatMap(lambda line: line.split())    .flatMap(lambda line: line.split())
    .map(lambda token: (token, 1))    .map(lambda token: (token, 1))
-   .reduceByKey(lambda x,y: x + y) +   .reduceByKey(lambda x, y: x + y) 
-   .sortBy(lambda (word,count)count, ascending=False)+   .sortBy(lambda word_countword_count[1], ascending=False)
    .saveAsTextFile(output))    .saveAsTextFile(output))
 sc.stop() sc.stop()
 </file> </file>
  
-  * run ''word_count.py'' script inside existing Spark cluster (i.e., inside ''spark-qsub'' or ''spark-qrsh''), or start local Spark cluster using as many threads as there are cores if there is none:+  * run ''word_count.py'' script inside existing Spark cluster (i.e., inside ''spark-sbatch'' or ''spark-srun''), or start local Spark cluster using as many threads as there are cores if there is none:
   <file>spark-submit word_count.py /net/projects/spark-example-data/wiki-cs outdir</file>   <file>spark-submit word_count.py /net/projects/spark-example-data/wiki-cs outdir</file>
   * run ''word_count.py'' script with local Spark cluster using one thread:   * run ''word_count.py'' script with local Spark cluster using one thread:
   <file>MASTER=local spark-submit word_count.py /net/projects/spark-example-data/wiki-cs outdir</file>   <file>MASTER=local spark-submit word_count.py /net/projects/spark-example-data/wiki-cs outdir</file>
-  * start Spark cluster (10 machines, 1GB RAM each) on SGE and run ''word_count.py'' script: +  * start Spark cluster (10 machines, @GB RAM each) using Slurm and run ''word_count.py'' script: 
-  <file>spark-qsub 10 1G spark-submit word_count.py /net/projects/spark-example-data/wiki-cs outdir</file>+  <file>spark-sbatch 10 2G spark-submit word_count.py /net/projects/spark-example-data/wiki-cs outdir</file>
  
 ===== Using Virtual Environments ===== ===== Using Virtual Environments =====

[ Back to the navigation ] [ Back to the content ]