Partager via

Output to Files (U-SQL)

  • Article

Summary

Currently U-SQL scripts are executed in batch mode and thus always need to persist their results either in a table or in a file. U-SQL’s OUTPUT statement writes the specified rowset to the provided file path URI. The file path URI is specified either as a relative or absolute Azure Data Lake Storage file path URI, a Windows Azure Blob Storage file path URI or a simplified path pattern expression.

  • Since file paths are URIs, characters such as spaces need to be URI-encoded. For example the file name a has 3 spaces has to be written as a%20has%203%20spaces.

  • A file path URI has to be provided as a static string expression (an expression that can be constant folded) in the TO clause of an OUTPUT expression.

  • A file path can also be specified as a local file path.

A local Azure Data Lake Storage file path (without the URI header) refers to a path in the Azure Data Lake Analytics account’s associated default Data Lake account.

The following grammar rules shows the syntax for each of the supported paths:

Syntax

Output_File_Path :=                                                                                      
     Default_Account_Path_URI 
|    ADL_Path_URI 
|    WASB_Path_URI 
|    Simple_File_Set_Path.

The grammar rules Default_Account_Path_URI, ADL_Path_URI and WASB_Path_URI are specified in Input Files (U-SQL).

Output File Sets

Unlike the input file paths, the output file path only supports a simplified file set path that can contain a single unnamed wildcard. The grammar rule Simple_File_Set_Path is defined as:

Syntax

Simple_File_Set_Path :=                                                                                  
     [ADL_Header | WebHDFS_Header | WASB_Header]+Simple_FileSet_Pattern.

Simple_FileSet_Pattern := 
     '/'+{directory_name +'/'}+ file_name | Simple_File_Name_Pattern.

Simple_File_Name_Pattern := 
     [{character}] Wildcard_Pattern [{character}].

Wildcard_Pattern := 
     '{*}'.

If a Simple_File_Set_Path is specified in an OUTPUT statement, the statement will create as many files as there were processing nodes that were processing the output rowset. Each file name will contain the node’s id in place of the Wildcard_Pattern.

Examples

  • The examples can be executed in Visual Studio with the Azure Data Lake Tools plug-in.
  • The examples below uses the sample data provided with your Data Lake Analytics account. See Prepare source data for additional information.
  • The sample data is used to create and populate the table defined below:
CREATE DATABASE IF NOT EXISTS TestReferenceDB;
USE DATABASE TestReferenceDB; 

DROP TABLE IF EXISTS dbo.SearchLog;
CREATE TABLE dbo.SearchLog (
       INDEX sl_idx CLUSTERED (UserId ASC) 
       DISTRIBUTED BY HASH (UserId) INTO 2
) AS  
    EXTRACT UserId          int  
          , Start           DateTime  
          , Region          string  
          , Query           string  
          , Duration        int  
          , Urls            string  
          , ClickedUrls     string
    FROM "/Samples/Data/SearchLog.tsv"
    USING Extractors.Tsv();

Default_Account_Path_URI

@result = 
    SELECT *
    FROM TestReferenceDB.dbo.SearchLog;

OUTPUT @result
TO "/Samples/Outputs/ReferenceGuide/OutA.csv" 
USING Outputters.Csv();

ADL_Path_URI

@result =
    SELECT *
    FROM TestReferenceDB.dbo.SearchLog;

OUTPUT @result
TO "adl://<adl_accountname>.azuredatalakestore.net/Samples/Outputs/ReferenceGuide/OutB.csv"
USING Outputters.Csv();

WASB_Path_URI

@result =
    SELECT *
    FROM TestReferenceDB.dbo.SearchLog;

OUTPUT @result
TO "wasb://<wasb_container>@<wasb_accountname>/ReferenceGuide/OutC.csv"
USING Outputters.Csv();

Simple_File_Set_Path

@result =
   SELECT *
   FROM TestReferenceDB.dbo.SearchLog;

OUTPUT @result
TO "/Samples/Outputs/ReferenceGuide/Out{*}.csv"
USING Outputters.Csv();

See Also