Dear Wiki user,

You have subscribed to a wiki page or wiki category on "Pig Wiki" for change 
notification.

The following page has been changed by OlgaN:
http://wiki.apache.org/pig/PigStreamingFunctionalSpec

------------------------------------------------------------------------------
  
  ==== 3.4 Ability to write output(s) to the disk ====
  
- The motivation here is similar for ability to read data from disk described 
in 3.3:
+ The motivation here is similar to the ability to read data from disk 
described in 3.3:
  
   1. To accommodate existing applications that write output to the disk rather 
than stdout.
   2. To provide ability to create multiple outputs.
@@ -219, +219 @@

  Y = stream X through Z;
  }}}
  
- This tells pig that streaming application stored its complete output into 
file called _outputfile_ and that the content of that file should be serialized 
into Y using !MySerializer. (The `outputfile` is written by `stream.pl` by 
using standard open/write/close functions. Pig might need to capture the data 
and store them as part files in DFS. )
+ This tells pig that streaming application stored its complete output into 
file called `outputfile` in the tasks's working directory and that the content 
of that file should be serialized into Y using !MySerializer. 
  
  A user can specify multiple outputs but only the first one will be 
automatically loaded; the rest would be stored in dfs using the file name 
specified in the output as absolute path:
  
@@ -228, +228 @@

  Y = stream X through A;
  }}}
  
- In this example, output1 would loaded into Y while the second output will be 
stored in DFS in the file called  '/data/mydata/output2'.
+ In this example, `output1` will be loaded into Y while the second output will 
be stored in DFS in the file called  '/data/mydata/output2'.
  
  [[Anchor(Non-streaming_Features)]]
  === 4 Non-streaming Features ===

Reply via email to