Over the last few months we have gotten the following question from a number of different sources;
"What's the difference between Oracle's various columnar technologies and how do they fit together"?
If you are struggling with the same question, we have the answer for you.
[Read More]The Database In-Memory team has been busy attending conferences in the
last couple of months letting you know what's new with Database
In-Memory. These conferences have included Open World, the East Coast
Oracle Users Conference (ECO), the Bulgarian Oracle User Group (BGOUG)
conference, the German Oracle User Group (DOAG) conference, the All
India Oracle Users Group (AIOUG) conference SANGAM16 and we would like
to thank everyone for attending our sessions and stopping by to ask
questions.
One
of the questions that continues to get asked is about Patch Set Updates
(PSUs). We're still finding that people are installing PSUs (see MOS
Note 854428.1), or even Critical Patch Updates (CPUs), and expecting to get fixes for Database In-Memory. We've blogged about this in the past here, but I think it's worth repeating.
[Read More]
With the introduction of Oracle Database Exadata Express Cloud Service I thought it would be nice to review where Database In-Memory is offered in Oracle's cloud services. The Oracle Cloud not only offers customers an easy and cost effective way to deploy their applications in the cloud, but it also allows customers to easily try out Oracle options that they might not have otherwise licensed. In the case of Oracle Database Exadata Express Cloud Service it is possible to not only try out Oracle Database 12.2 but also the new 12.2 features of Database In-Memory with the X50IM option. The Oracle Database Exadata Express Cloud Service can be an excellent entry level environment for a test or development system.
Oracle offers three database cloud services that have options for Database In-Memory:
Currently the Oracle Database Cloud Service and Oracle Database Exadata Cloud Service options support Database In-Memory on Oracle Database 12.1.0.2. The Oracle Database Exadata Express Cloud Service supports Database In-Memory on Oracle Database 12.2 with the XIM50 option.
[Read More]
There is less than 6 weeks to go until Oracle Open World, the largest gathering of Oracle customers, partners, developers, and technology enthusiasts, which begins on September 18th in San Francisco.
Of course the In-Memory development group will be there and you will have multiple opportunities to meet up with us, in one of our technical sessions, our hands-on-labs or at the Oracle demogrounds.
This year the In-Memory team has 4 technical sessions and there are also a number of excellent customer sessions you shouldn't miss.
[Read More]
A new version of the Oracle Database In-Memory Advisor has been released. The In-Memory Advisor is designed to help identify analytic processing that may benefit from the use of Oracle Database In-Memory. It supports Oracle Database 11.2.0.3 and later so you don't have to upgrade to Oracle Database 12c to try it out. The new version includes a number of enhancements including an enhanced and simplified script to run the Advisor and get recommendations.
You can get more information about the In-Memory Advisor here.
A couple of weeks ago at the Spring Conference of the Bulgarian Oracle User Group, Joel Goodman gave a great session on Bloom Filters and how the can improve the performance of hash joins in the Oracle Database especially in the In-Memory column store. At the end of that session the chap sitting in front of me turned around to show me an execution plan they got while testing Oracle Database In-Memory and posed the following question, “Where did the other bloom filter go?”
[Read More]We tell people that Database In-Memory is all about improving analytic
processing. The definition that I use is "using aggregation to find
patterns or trends in data". I believe I saw this used in one of Juan Loaiza's
OpenWorld presentations and it made sense to me. After all Database
In-Memory is really good at scanning and filtering a lot of data, and
most of the articles on this blog echo that theme.
The problem
is, the definition that I use doesn't necessarily make sense for
everyone. When we do presentations or give webinars many people don't
seem to really get what we mean by analytics.
[Read More]
I was recently asked by a customer, "is it possible to force a query to use the In-Memory column store?" and I thought it would be a good idea to share the answer here on the blog.
The first thing that comes to mind when you starting thinking about forcing how the Oracle Database will execute a query is Optimizer hints. There is an INMEMORY hint available, but that hint just enables the use of the In-Memory column store (IM column store) if the INMEMORY_QUERY parameter has been set to DISABLE (default is ENABLE). If we look closer at the documentation for the hint, we see that it says "This hint does not instruct the optimizer to perform a full table scan. If a full table scan is desired, then also specify the "FULL hint".
So it appears that the answer is that you have to use the FULL hint to ensure that the access method is a full table scan which is a requirement for accessing the IM column store. If the INMEMORY_QUERY has been set to DISABLE then you would also need to use the INMEMORY hint. Of course this also assumes that the IM column store has been enabled and that the object has been enabled for in-memory.
[Read More]
The compression techniques that are use for in-memory compression are different from what we use for on-disk compression. The compression on-disk is designed to reduce the overall space usage for the table segments. While the compression algorithms used in memory are designed for query performance and speeding up the queries by being able to operate on the compressed format. Due to the different techniques, it is not uncommon to have table segments see different compression ratios and different sizes in-memory and on-disk.
[Read More]We've added a new Technical Briefs section to the blog. If you scroll down on the blog you will see a new section titled "Technical Briefs" on the right hand side.
One of the briefs is an update for Oracle Database 12c on how to create SQL Monitor active reports and the other describes how to create an Optimizer trace file. Hopefully you will find these useful quick references.
I recently got a question from a customer who has just begun adopting Database In-Memory and thought it was worth answering as a blog post in case anyone else had encountered the same scenario and was wondering what was going on.
Here is the question I got:
“I've observed a scenario where a table is fully populated in the In-Memory column store with 0 in the BYTES_NOT_POPULATED column in v$im_segments, but after some time, the BYTES_NOT_POPULATED column shows a non-zero value. Why is Oracle purging part of my table from memory?”
Let me begin to answer this by assuring you all that Oracle does not purge your data out of the In-Memory column store.
[Read More]After I posted last weeks blog post on identifying analytic queries that benefit from Database In-Memory, I got the following question, Why didn't Database In-Memory improve the performance of my query?"
I've simplified the query in question down to the following,
SELECT Count(DISTINCT cust_id)
FROM ssh.sales
WHERE amount_sold > 1;
So, why didn’t Database In-Memory improve the performance of this query?
[Read More]Although Database In-Memory can be used in both Enterprise OLTP systems and Data Warehouses, only analytic queries will benefit from accessing data from the In-Memory column store (IM column store). Remember the IM column store enables large volumes of data to be rapidly scanned and processed using a variety of optimizations. Simpler transaction based queries, such as a primary key look up do not benefit from the IM column store and are automatically directed to the buffer cache.
But how do you identify an analytic query?
An analytic query is one where a large amount of data is scanned and filtered to return critical data that drives better business decisions. It doesn’t necessarily mean a query that contains one of Oracle’s analytical functions but they can also see benefits. A star query is a good example of such a query. By star query, I mean a query where a large fact table (center of the star) is joined to a number of smaller dimension tables (the points of the star).
If you don’t have a simple star schema (most people don’t) then you can identify a good candidate query by examining some key characteristics.
[Read More]Andy Rivenes (@TheInMemoryGuy) from the Oracle Database In-Memory team will be presenting two hands-on-labs and will be co-presenting a session at this year's IOUG Collaborate 16 conference, in Las Vegas April 10 - 14. Below are details on the sessions he will be presenting. We hope that you have an opportunity to check out some of these sessions if you plan to attend the event!
[Read More]
The Oracle Database In-Memory blog is written by the Oracle product management team and sheds light on all things In-Memory.
