Oracle GoldenGate 23ai: Parameters, Data Selection, Filtering, & Transformation
Description
Welcome to the Oracle University Podcast, the first stop on your cloud journey. During this series of informative podcasts, we’ll bring you foundational training on the most popular Oracle technologies. Let’s get started!
Lois: Hello and welcome to the Oracle University Podcast! I’m Lois Houston, Director of Innovation Programs with Oracle University, and with me is Nikita Abraham, Team Lead: Editorial Services.
Nikita: Hi everyone! This is the last episode in our Oracle GoldenGate 23ai series. Previously, we looked at how you can manage Extract Trails and Files. If you missed that episode, do go back and give it a listen.
Lois: Today, Nick Wagner, Senior Director of Product Management for GoldenGate, is back on the podcast to tell us about parameters, data selection, filtering, and transformation. These are key components of GoldenGate because they allow us to control what data is replicated, how it's transformed, and where it's sent. Hi Nick! Thanks for joining us again. So, what are the different types of parameter files?
Nick: We have a GLOBALS parameter file and your runtime parameter files. The global one is going to affect all processes within a deployment. It's going to be things like where's your checkpoint table located in name, things like the heartbeat table. You want to have a single one of these across your entire deployment, so it makes sense to keep it within a single file.
We also have runtime parameter files. This are going to be associated with a specific extract or replicat process. These files are located in your OGG_ETC_HOME/conf/ogg. The GLOBALS file is just simply named GLOBALS and all capitals, and your parameter file names for the processes themselves are named with the process.prm.
So if my extract process is EXT demo, my parameter file name will be extdemo.prm. When you make changes to parameter files, they don't take effect until the process is restarted. So in the case of a GLOBALS parameter file, you need to restart the administration service. And in a runtime parameter file, you need to restart that specific process before any changes will take effect.
We also have what we call a managed process setting profile. And this allows you to set up auto restart profiles for each process. And the GoldenGate Gate classic architecture, this was contained within the GLOBALS parameter file and handled by the manager. And microservices is a little bit different, it's handled by the service manager itself. But now we actually set up profiles.
Nikita: Ok, so what can you tell us about the extract parameter file specifically?
Nick: There's a couple things within the extract parameter file is common use. First, we want to tell what the group name is. So in this case, it would be our extract name. We need to put in information on where the extract process is going to be writing the data it captures to and that would be our trail files, and extract process can write to one or more trail files.
We also want to list out the list of tables and schemas that we're going to be capturing, as well as any kind of DDL changes. If we're doing an initial load, we want to set up the SQL predicate to determine which tables are being captured and put a WHERE clause on those to speed up performance. We can also do filtering within the extract process as well. So we write just the information that we need to the trail file.
Nikita: And what are the common parameters within an extract process?
Nick: There are a couple of common parameters within your extract process. We have table to list out the list of tables that GoldenGate is going to be capturing from. These can be wildcarded. So I can simply do table.star and GoldenGate will capture all the tables in that database. I can also do schema.star and it will capture all the tables within a schema. We have our EXTTRAIL command, which tells GoldenGate which trail to write to.
If I want to filter out certain rows and columns, I can use the filter cols and cols except parameter. GoldenGate can also capture sequence changes. So we would use the sequence parameter. And then we can also set some high-level database options for GoldenGate that affect all the tables and that's configured using the tranlog options parameter.
Lois: Nick, can you talk a bit about the different types of tranlogoptions settings? How can they be used to control what the extract process does?
Nick: So one of the first ones is ExcludeTag. So GoldenGate has the ability to exclude tagged transactions. Within the database itself, you can actually specify a transaction to be tagged using a DBMS set tag option.
GoldenGate replicat also sets its transactions with a tag so that the GoldenGate process knows which transactions were done by the replicat and it can exclude them automatically. You can do exclude tag with a plus. That simply means to exclude any transaction that's been tagged with any value. You can also exclude specific tags.
Another good option for TranLogOptions is enable procedural replication. This allows GoldenGate to actually capture and replicate database procedure calls, and this would be things like DBMS AQ, NQ operations, or DQ operations. So if you're using Oracle advanced queuing and you need GoldenGate to replicate those changes, it can.
Another valuable tranlogoption setting is enable auto capture. Within the Oracle Database, you can actually set ALTER TABLE command that says ALTER TABLE, enable logical replication. Or when you create a table, you can actually do CREATE TABLE statement and at the end use the enable logical replication option for that CREATE TABLE statement. And this tells GoldenGate to automatically capture that table.
One of the nice features about this is that I don't need to specify that table and my parameter file, and it'll automatically enable supplemental logging on that table for me using scheduling columns. So it makes it very easy to set up replication between Oracle databases.
Nikita: Can you tell us about replicat parameters, Nick?
Nick: Within a replicat, we'll have the group name, some common other parameters that we'll use is a mapping parameter that allows us to map the source to target table relationships. We can do transformation within the replicat, as well as error handling and controlling group operations to improve performance.
Some common replicat parameters include the replicat parameter itself, which tells us what the name of that replicat is. We have our map statement, which allows us to map a source object to a target object. We have things like rep error that control how to handle errors. Insert all records allows us to change and convert, update, and delete operations into inserts. We can do things like compare calls, which helps with active-active replication in determining which columns are used in the GoldenGate WHERE clause. We also have the ability to use macros and column mapping to do additional transformation and make the parameter file look elegant.
AI is being used in nearly every industry…healthcare, manufacturing, retail, customer service, transportation, agriculture, you name it! And it’s only going to get more prevalent and transformational in the future. It’s no wonder that AI skills are the most sought-after by employers. If you’re ready to dive in to AI, check out the OCI AI Foundations training and certification that’s available for free! It’s the perfect starting point to build your AI knowledge. So, get going! Head on over to mylearn.oracle.com to find out more.
Nikita: Welcome back! Let’s move on to some of the most interesting topics within GoldenGate… data mapping, selection, and transformation. As I understand, users can do pretty cool things with GoldenGate. So Nick, let’s start with how GoldenGate can manipulate, change, and map data between two different databases.
Nick: The map statement within a Replicat parameter allows you to provide specifications on how you're going to map source and target objects. You can also use a map and an extract, but it's pretty rare. And that would be used if you needed to write the object name. Inside the trail files is a different name than the actual object name that you're capturing from.
GoldenGate can also do diffe
United States
