Schema Mapping

Schema mapping means matching the field names and formats of data coming from different sources, marking them as "these mean the same thing" so the data can be combined into a single table. For example, a ticketing company might call a customer's name customer_name, while a streaming service calls the same information user_full_name. These two fields have different names and different storage formats, but to combine the data from both systems, you need to mark that they actually represent the same piece of information.

This is similar to sorting moving boxes from several different households: each box has a different label, so you have to look at them and re-sort them into categories like "kitchen items" or "bedroom items" to organize everything in one place. Every time a company added a new data source, someone had to do this matching by hand, which often meant it took weeks to connect just one source. Recently, more efforts have focused on getting AI to handle this matching process automatically.

In articles, it shows up in phrasing like "schema mapping, pipeline building, data quality checks, and GDPR classification work were all done manually." A common misunderstanding is that schema mapping changes the data content itself. It doesn't modify values — it creates a map connecting fields that are stored under different names and formats but actually carry the same meaning.

You can ask a chatbot the following to get a feel for what schema mapping involves.

Match up the fields between these two tables. Table A has customer_name, signup_date, and phone_no. Table B has user_full_name, registration_date, and contact_number. Determine which fields mean the same thing and organize the result into a mapping table.