Before you begin with the migration process, the following steps should be completed.

• Your account should be provisioned with access to Personalization. 
  
• Your account should be configured with an Acoustic Exchange account. This will enable forwarding of the Legacy Digital Recommendation events to the Personalization components for processing.
• Your currently configured channels (websites) will be migrated to leverage the new Personalization capabilities. 
  
• Legacy Digital Recommendations offers will be migrated to product recommendation strategies.

Product catalogs

Legacy Digital Recommendation catalog files should be uploaded to a new endpoint using the same file format that is currently being used. The following sections provide guidelines for each type of file.

 EPCMF file

• This is the main Catalog file in CSV format.
• The file must include EPCMF keyword in the file name.
• These are the mandatory fields in the EPCMF file and they should be specified in the same sequence as shown here: Date ID, Client ID, Product ID, Product Name
  Note: The field names can vary based on columns.txt
• The sequence of columns in EPCMF file must match the sequence of columns specified in the columns.txt file.

ECDF file (Optional)

• This is Category Hierarchy file in csv format.
• The file must include ECDF keyword in file name.
• All the fields in the ECDF file are mandatory, and they should be specified in the same sequence as shown here: Date Id, Client Id, Category Id, Category Name, Parent Category Id

Headers file

• File name can either be columns.txt or columns.csv.
• This file has a specific sequence of content, as explained below:
• 
  

  Line #	Description
  1	Specify the column names.  The total number of attribute names in columns.txt should match the total number of columns in EPCMF file. Extra columns from columns.txt will be ignored if not sent in EPCMF file. Lesser number of attributes in columns.txt than EPCMF columns will result in an error. The sequence of attributes in columns.txt should match the sequence of columns in EPCMF file. Empty columns in EPCMF should have associated empty attributes in columns.txt. Here's an example for pipe-separated file: Expected EPCMF data format (Extra pipes ('|') indicate empty columns): 20200721|53270000|B-AC00182-19617|Womens Sailor Socks Mini Cornish Peacock Bass|5637145545|6|0|Blue|PROMO|SS20-L2|No|AC00182||||||||||||| Associated columns.txt headers (Extra commas indicate empty columns) : Date ID,Client ID,Product ID, Product Name,Category,Price,Inventory,Colour,Price Status,Drop ID,Plus,Style ID,Product URL,Product Image URL,,,,,,,,,,,Temp Product Family To update an existing record, ensure that the attribute name in the new upload is same as the one in the original file. Otherwise, they are treated as different records.
  2	Specify the datatype for each column.  Datatypes should match the number of columns. The datatypes you can specify in this file are: Text and Number. Default data type for empty values is Text.
  3	Specify the delimiter to be used in the EPCMF & ECDF files.  The following delimiters are supported: comma  OR pipe(|) OR  tab
  4	Provide a numeric value that indicates the index position of the ImageURLAttr attribute. If you do not require ImageURLAttr attribute, specify the value as -1.
  5	Provide a numeric value that indicates the index position of the PriceAttr attribute. If you do not require PriceAttr attribute, specify the value as -1.
  6	For ImageURLPrefix, provide the base URL for the image URL to form the complete image link. Do not include the protocol. Example: //img.citycool.com/is/image/graphics/ If you provided absolute path for the image in the image_url field, then for ImageURLPrefix field, specify the value as  -1.
  7	Specify whether to process the ECDF file or not (ECDF = true OR false). If you specify this value as true, and the ECDF file is not available, the product catalog upload will fail. Similarly, if you specify this value as false, the ECDF file is not processed even if it is available.

  Here's an example of columns.txt file: 
  
  

Notes

• Do not include the headers/column names in EPCMF & ECDF files. The headers are picked up from  columns.txt file. Any headers included in EPCMF & ECDF files will be ignored.
• If you need to send multiple values in a single column:
  • You must use pipe(|) as the separator between the multiple values within a column.
  • In this case, the delimiter field value cannot be pipe in columns.txt file.

Adding custom attributes

If you want to use custom attributes, you should include the attribute names along with the datatype  in the columns.txt file. Values for these custom attributes should be present in EPCMF file in the respective  columns as per the columns.txt file.