Important Considerations for All Files

Double quotes

All data sent should be wrapped in double quotes:

• Actual quotation marks should be denoted with two quotation marks. For example,  """Hello"" World" to send "Hello" World value.
• Blank fields should be sent as blank, not as "". Double quotation marks ("") are interpreted as an empty string rather than a null value.

Record IDs

• It is critical for record_id's to be consistent across all flat files. It is how we know things like categories/items are linked together.
• record_id's for Flat Files must match the id's used in the JavaScript. For example, if a JS cart wave has an item id of “123”, an item with that record_id must exist in the “Items” table.

UTC DateTime vs. UTC Date

UTC DateTime vs. UTC Date is an important distinction/ For example, if a field is listed as DateTime and only Date is provided, that row is flagged as invalid and won't be imported.

Optional fields

While many fields are Optional, if the data is readily available it is highly recommended to provide to RS the maximum data for AI to leverage in decision making.

No data for a column

If there is no data for a column, make sure the field comes over as blank. For example, “NULL”, "0", “NO”, or "", can be interpreted as strings.

Field impact on Cortex performance

Each field provided has some type of impact on the final campaign performance of Cortex. These impacts are discussed in the “Context/Considerations” column found in the file schemas article, but please contact your technical onboarding contact at ReSci if you have any additional questions.

Parent_record_id columns

Remember that the parent_record_id referenced in a row should always match another row’s record_id. For example, parent items must exist as their own standalone row/entry.