Converting Oracle data types to PostgreSQL

You're an expert in Oracle. >> [music] >> You know varchar2, number, and date like the back of your hand. When you move to Postgres, subtle differences in data types can cause major headaches. What if your migration tool could not only convert your schema, but also be your personal type tutor? I'm going to show how Google Cloud's Database Migration Service does exactly [music] that. In my DMS conversion workspace, I have an Oracle source connected. Looking at this employees table, on the left, [music] we have the original Oracle DDL, and on the right, we have the automatically converted code for Cloud SQL for Postgres. >> [music] >> So, check it out. Last name, varchar2, was converted by DMS to last name, varchar. [music] But, why? Let's ask Gemini. I'll click on the converted line and ask for an explanation. [music] Gemini explains to me that varchar2 is an Oracle specific type, >> [music] >> but it's been mapped to the standard SQL varchar. Okay, it's good. But, it also points out that in [music] Oracle, the length could mean bytes or characters, but in Postgres, it always [music] means characters, which simplifies handling multi-byte languages. DMS just gave us a quick internationalization lesson. >> [music] >> Okay, next up, salary number becomes salary decimal. Let's see [music] why. Gemini explains that Oracle's all-purpose number type is [music] mapped to Postgres's decimal to preserve the exact precision and scale. This is critical for financial data. It also notes that for integer-only numbers, DMS might suggest bigint for better performance. Another best practice we learned on the fly. Finally, let's look at hire date, date, [music] becomes hire date, timestamp without time zone. Gemini clarifies that Oracle's date type [music] actually stores both date and time components down to the second. To maintain that level of precision, DMS maps is Postgres's [music] timestamp type, which does the same. This prevents silent truncation of time data, [music] a common migration pitfall. With DMS, you aren't just blindly accepting an automated conversion. You're engaging with it, learning from it, and building trust. This process reduces cognitive load and ensures your new database is built on a solid, correct foundation from day one. >> [music] >> Are you ready to convert your schema? Explore the Database Migration Service Conversion Workspace today. The link is in the description below. >> [music]