Just a pointer to two blog posts that I find worth mentioning:
1. Christo Kutrovsky from Pythian writes about some quirks he found regarding Parallel Distribution of aggregation and analytic functions. In particular the lower part of the post (which is not about the initial Interval Partitioning issue) gives a lot of food for thought how the chosen Parallel Distribution can influence the performance of operations
2. Alexander Anokhin gives some geek insights into how Oracle performs logical I/O and in particular how the "buffer is pinned count" statistics is maintained by Oracle. He even introduces another cool tool ("Digger") for performing such investigations based on DTrace.
That post quite nicely complements my "Logical I/O evolution" series.
Saturday, July 28, 2012
Sunday, July 22, 2012
Free Webinar
In a couple of days, on Wednesday, 1st of August, I'll be presenting another free webinar hosted at AllThingsOracle.com.
Although it is called "Oracle Cost-Based Optimizer Advanced Session", don't be mislead by the title.
It is not a truly "advanced" session, but rather I'll try to delve into various topics that I could only mention briefly or had to omit completely during the first webinar on the Cost-Based Optimizer.
In principle it's going to be a selection of the most recurring issues that I come across during my consultancy work:
- I'm going to spend some time on statistics and histograms in particular and what I believe are the most important aspects to understand regarding them
- Why the clustering factor of indexes gathered by the statistics collection might be wrong (this was only briefly mentioned in the first webinar)
- Using the wrong datatype and how this impacts the estimates of the optimizer is going to be another potential topic
- Implicit datatype conversions are also an issue I come across quite frequently (Tom Kyte covered this already to some extent in one of the previous AllThingsOracle.com webinars I believe)
- Last but not least Dynamic Sampling, Virtual Columns and Extended Statistics are also worth to talk about if time permits it
This time I'm going to use only a couple of slides and mainly will run through some test cases that hopefully allow a live demonstration of what I'm talking about.
You can find more details and register here.
Although it is called "Oracle Cost-Based Optimizer Advanced Session", don't be mislead by the title.
It is not a truly "advanced" session, but rather I'll try to delve into various topics that I could only mention briefly or had to omit completely during the first webinar on the Cost-Based Optimizer.
In principle it's going to be a selection of the most recurring issues that I come across during my consultancy work:
- I'm going to spend some time on statistics and histograms in particular and what I believe are the most important aspects to understand regarding them
- Why the clustering factor of indexes gathered by the statistics collection might be wrong (this was only briefly mentioned in the first webinar)
- Using the wrong datatype and how this impacts the estimates of the optimizer is going to be another potential topic
- Implicit datatype conversions are also an issue I come across quite frequently (Tom Kyte covered this already to some extent in one of the previous AllThingsOracle.com webinars I believe)
- Last but not least Dynamic Sampling, Virtual Columns and Extended Statistics are also worth to talk about if time permits it
This time I'm going to use only a couple of slides and mainly will run through some test cases that hopefully allow a live demonstration of what I'm talking about.
You can find more details and register here.
Subscribe to:
Posts (Atom)
