Scenario 2: Standardizing addresses from unstructured data
This scenario applies only to a subscription-based Talend Platform solution or Talend Data Fabric.
In this scenario, six components are used to standardize addresses from unstructured
input data copied from a Website, by matching the data with the data contained in an
index previously generated by a Job.
For more information about index creation, see tSynonymOutput.
Drop the following components from the Palette to the
tFixedFlowInput: this component holds the
unstructured data from which the addresses are extracted.
tStandardizeRow: this component defines the
address rules and generates the addresses in XML format with the defined tags.
This is the process of normalizing and standardizing the initially unstructured
tFilterColumns: this component filters the
tExtractXMLField: this component extracts
the attributes from the Address node of the XML tree in
order to output every address item in formatted columns.
two tLogRow: these components are used to
display the output data. The first tLogRow
returns the errors, if any. The second tLogRow
displays the result in the console.
Before starting up to replicate this scenario, you have to retrieve the content of an
index in order to match the unstructured data with the index data. The content of the
index reads as follows:
On the left side are held Paris street names and on the right side are held the
synonyms used in the data. The data will be used as references to standardize the
address data collected from the website of interest, for example, from http://paris.conciergerie.com/travel_paris/address.php.
To replicate this scenario, proceed as the following sections illustrate.