Scenario: Generating duplicate data from an input flow - 6.1

Talend Components Reference Guide

Talend Big Data
Talend Big Data Platform
Talend Data Fabric
Talend Data Integration
Talend Data Management Platform
Talend Data Services Platform
Talend ESB
Talend MDM Platform
Talend Open Studio for Big Data
Talend Open Studio for Data Integration
Talend Open Studio for Data Quality
Talend Open Studio for ESB
Talend Open Studio for MDM
Talend Real-Time Big Data Platform
Data Governance
Data Quality and Preparation
Design and Development
Talend Studio

This scenario describes a basic Job that generates a sample of duplicate data from an input flow by using probability theories and specific criteria on three columns: Name, City and DOB (date of birth).

Below is a capture of a sample data of the input flow:

Setting up the Job

  1. Drop the following components from the Palette onto the design workspace: tFileInputDelimited, tDuplicateRow and tFileOutputDelimited.

  2. Connect all the components together using the Row > Main link.

Configuring the input data

  1. Double-click tFileInputDelimited to display the Basic settings view and define the component properties.

  2. In the File name/Stream field, browse to the file to be used as the main input.

    This file provides some information about customers.

  3. Define the row and field separators the header and footer in the corresponding fields, if any.

  4. Click the [...] button next to Edit schema to open a dialog box and define the input schema. According to the input file structure, the schema is made of ten columns.

  5. Click the [+] button and define the input columns in the dialog box as in the above figure. Click OK to close the dialog box.

  6. If needed, right-click tFileInputDelimited and select Data Viewer to display a view of the input data.

Configuring the duplicate data

  1. Double-click tDuplicateRow to display the Basic settings view and define the component properties.

  2. Click the Edit schema button to view the input and output columns and do any modifications in the output schema, if needed.

    The output schema of this component contains one read-only column, ORIGINAL_MARK. This column identifies, by true or false, if the record is an original or a duplicate record. There is only one original record per group of duplicates.

  3. In the Percentage of duplicated records field, enter the percentage of the duplicate rows you want to have in the output flow.

  4. In the Distribution of duplicates area, select the Bernoulli distribution (probability theory) you want to use to generate duplicates. Set an average of how many duplicate records to have in each group.

  5. Click the plus button below the Modifications table and add four lines in the table.

    This table enables you to define what values to change in a given column and how to change them in order to generate duplicates:

    1. In the Input Column, select the column from the input flow from which you want to generate duplicates, Name, City and DOB in this example.

      When you add a column twice in the table and select different functions, you generate duplicates from the same field with different values. For example, in this scenario you modify the duplicate names with the Soundex replace function with probability of 50%, and then you modify it again with the Exchange characters function with the probability of 50%. This means, the Name field of the duplicate record may still not be modified after the second function with the following probability: (1-0.5)*(1-0.5) = 0.25.

      If you want to make sure that all the duplicated records are modified, you must set the Modification Rate to 1.

    2. In the Modification Rate column, enter a rate of the duplicate records you want to generate from the input column.

    3. From the Function list, select the function that will decide what modification to do on a value to duplicate it.

      In this example, there will be duplicate names with different sounds and characters and duplicate city names with different sounds. Date values in the date of birth column will be randomly changed here.

    4. In the Max Modification Count column, enter a maximum number of the values to be modified in each field.

  6. Click the Advanced settings tab and enter a random number in the Seed for random generator field.

    By setting a number in this field, you will generate the same sample of duplicate data in each execution of the Job. Change the value if you want to generate a diffierent sample.

Configuring the output component

  1. Double-click tFileOutputDelimited to display its Basic settings view and define the component properties.

  2. In the File Name field, specify the path to the file to which you want to write the duplicate data, duplicated_records in this example.

  3. Define the row and field separators in the corresponding fields, if any.

Executing the Job

  1. Save your Job and press F6 to execute it.

    Duplicate data is generated and written to the output file.

  2. Right-click the output component and select Data Viewer to display the duplicate data.

    Duplicate records have been marked as false in the ORIGINAL_MARK column.

    Some data has been modified in the Name, City and DOB fields according to the criteria you set in the Modifications table and duplicate records have been generated based on these modifications.

    For example, if you compare the original name Mrs Morgan Ross and the duplicate name Mrs M rganosRiss, you will see that the two functions have been used on this duplicate record: the letter o has been exchanged with a space, and also the sound has been replaced in Ross and Riss. However, the soundex code has not been changed for the replaced sound.

  3. In the tDuplicateRow basic settings and in the Distribution of duplicates area, select a different distribution, Bernoulli distribution for example, and run the Job.

    Different duplicates are generated from the same input flow according to the selected distribution as shown in the below figure.