AdminStephen Kwan
(Product Manager, OSIsoft)
My feedback
-
38 votes
An error occurred while saving the comment An error occurred while saving the comment In calculating your UnitStatus, do you use PI Points in your Analysis?
An error occurred while saving the comment You can do that with currently shipping software. Just configure an attribute with a Formula Data Reference and a %StartTime% or %EndTime% substitution parameter.
This would not work if you wish to write the start and end time to a PI Point.
If I have misinterpreted your needs, please let me know.
An error occurred while saving the comment Can you provide some example use cases?
-
2 votes
An error occurred while saving the comment Hi, can you tell me which tool you're using to create these event frames?
-
20 votes
An error occurred while saving the comment Hi, thanks for your suggestion and comment. In reviewing this suggestion, I have some questions that I hope you can help me with. What's the end goal of changing the event frame name? Is it to identify event frames that you can easily find? If that's the case, have you considered searching for event frames based on an attribute value? Setting the name of the event frame on close by PI Analysis Service has a side effect in that searches could have different results for in-progress event frames. For example, if an event frame was open, we have to give it a name. If we then subsequently change the name when the event frame closes, then any previous searches based on the name of the in-progress event frame would have different results. Any input you may have is appreciated.
-
2 votes
An error occurred while saving the comment Can you provide some examples of what you would like to do? Who would be configuring and applying these filters? What would an example of such filters?
-
3 votes
An error occurred while saving the comment Hi, can you help me understand what you're trying to do? Are you trying to uninstall AFLink because you don't need it?
-
1 vote
An error occurred while saving the comment There are some complexities associated with this. When creating objects from template, inevitably you would end up with many with the same name. As such, sorting would require the entire path. This would require careful design to ensure good user interface. In addition, paging becomes more difficult. Imagine you have 100k analyses. In order to sort or re-sort, we would have to load the entire collection, which is impactful to performance.
Having said that, we will look at what we can do.
-
3 votes
An error occurred while saving the comment I want to touch on some of the reasoning behind how we package (bundle). All these tools are reliant on the data access layer under the hood, for example AF SDK. As such, they are really clients from that perspective. They also have version dependencies on the SDK. Let's look at an example, PI System Explorer is nothing more than a shell on top of the AF SDK. As such, it is a client and must be bundled and shipped with the AF SDK as there is a version dependency for specific features. While we work very hard to support backwards compatibility, realistically for the best user experience, it has to be shipped (bundled) with the companion version of the AF SDK. This is the same with PI Builder. PI System Management tool requires PI SDK, AF SDK and other internal only components under the hood. Therefore, there are similar restrictions.
We provide an AFClient setup kit because we have 3rd party developers/partners that develop against the AF SDK. Therefore it's necessary to have a separate AFClient setup kit. This and what I've already describe above are the reasons why the AFClient kit includes PI System Explorer and PI Builder.
As to SMT bundling with ICU, it's more of a convenience for end users with some specific use cases. In SMT, you can configure certain interface points that under the hood require %OSI% in the Module Database. %OSI% modules are typically created when you configure interfaces with ICU. Thus, the decision was made long ago to bundle ICU with SMT.
We don't bundle PI DataLink with PI Builder because we sell PI DataLink as a separate product and for those who don't use AF or if they are not licensed for AF, it may not make sense to bundle PI Builder with PI DataLink.
As to confusion with which tool to do what, I would agree that's an area that we can improve on. As tools evolve, sometimes overlaps occur but we do try very hard to differentiate them. For example, PI Builder was designed to be a configuration tool so it doesn't really retrieve data in mass like PI DataLink. On thing that we can do is to start deprecating legacy products.
Having said all that, we can always re-examine whether there are better ways to rebundle.
-
6 votes
An error occurred while saving the comment The original design decision was not for PI Builder to replace DataLink because they serve two different purposes. As such, PI Builder was enhanced to replace the legacy PI Tag Configurator, which is for configuration only. Having said that, I can see how this suggestion may be useful in some cases.
Lastly note that you can retrieve the snapshot value of AF attributes that are configured as PI Point Data References. This may help in configuring other data references.
-
1 vote
An error occurred while saving the comment Unfortunately we actually don't know when a disconnection has occurred. Just because there is no data flowing, it doesn't always mean there is a disconnection. There is no "marker" that tells us a disconnection has occurred. Even if there was a marker, the fact that the connection is disconnected mean that we wouldn't receive that marker.
No Data is a System Digital State that has a special meaning, especially when the clients make a data access call to the server. We cannot place No Data into a data stream just because there has been no data, because as I described earlier, we don't actually know when a disconnection has occurred.
-
1 vote
An error occurred while saving the comment Jeff,
We append the %ID% because that's unique and is necessary when you work with templates. As an example, if you have an element template which includes an analysis template, and this analysis template creates tags because you're storing the calculation results. With that configuration, when you create X instances of the element from template, you would create X instances of the analyses. Without a %ID%, all these analyses would be writing their results to the same (one) output PI Point.
We encourage our users to not worry about the PI Point name, but rather work through AF attributes and if you do that, the fact that the PI Point name as an unique ID appended to the end is irrelevant.
-
1 vote
An error occurred while saving the comment The recommendation is to monitor the PI Analysis Service evaluation count performance counter. Does that work for your needs?
-
108 votesRESEARCHING / EVALUATING · 24 comments · PI Server » Asset Framework (AF) · Flag idea as inappropriate… · Admin →
An error occurred while saving the comment Hi Dan, can you give some examples of what you're looking to do? Your comment doesn't provide enough information for me to understand what problems you're trying to solve.
An error occurred while saving the comment In response to Jagdish Konathala, "Thanks Steve. We had issues upgrading to..."
Jagdish:
Yes, SP2a does fix the installation issue with SQL Server AlwaysOn Availability groups.An error occurred while saving the comment AF 2018 SP2a has a new feature for splitting strings. Please see release note:
144299 - Split function was added to Expression and Event Frame Generation analyses to split a string into a string array.
An error occurred while saving the comment In response to Chris Beaupre, "A String Length and Basic Delimiter func..."
Can you help me understand what you're trying to do with tabular alarm data? As you know, an AF data reference such as the string builder is only executed on demand by the client. In order to have a fruitful discussion, it would be great if we can get some details on your use cases.
--
Steve KwanAn error occurred while saving the comment In response to John Messinger, "Hi David, Some of the additional funct..."
John,
What are you trying to do? Are you parsing alarms and events?
--
Steve Kwan -
1 vote
An error occurred while saving the comment Can you provide your use case for this request? The AFTable cache interval does not automatically update, it only updates if the interval has expired AND there's user access. If there is no user access, even if the interval has expired, the table doesn't automatically update.
-
17 votes
An error occurred while saving the comment Unfortunately we have not yet made the decision to implement this request.
An error occurred while saving the comment If you were to historize the attribute name that had the max value, you would need to store that in a separate PI Point in addition to the values. That means you have to have a way to know that these 2 PI Points are related. Is that OK?
-
3 votes
An error occurred while saving the comment Thanks for the quick reply. So let's say you have an Expression Analysis, working from left to right in the UI, you create a Variable, configure some equation, then map the output to an attribute. Now if you move this output attribute, you want the software to recognize that and adjust accordingly? Can you confirm if I got that use case correctly?
An error occurred while saving the comment Hi, looking for a bit more info on this. Can you clarify what you mean by "if a variable is mapped to an attribute"? Are you referring to a variable in analytics? Or something else?
-
7 votes
An error occurred while saving the comment Hi, I'm trying to understand a bit more of what you're trying to do. You have a 6 member PI Collective which by design should have the same data fanned by BufSS. But you're describing a situation where Asset Analytics would be writing to 4 of those members, rather than all 6. Did I get that right? In that case, you would have different data in the collective members, which is very unusual. How would your clients be accessing the data? Would you configure the clients to connect to only specific collective members? This is very unusual.
If you're trying to load balance, perhaps what you need are two different collectives, each with the necessary number of collective members. You can then read from one collective and write to the other collective.
-
4 votes
An error occurred while saving the comment In the current implementation, AFTables are library objects, meaning all users have access. If we allow users who are not admins to create and edit linked tables, then that means other users of the AF system would be able to access these tables. Is that your use case?
-
1 vote
An error occurred while saving the comment The benefit of using an Output Attribute is that you can historize these values, but I can see how your suggest may be useful. As a matter of curiosity, if you had to create output attributes, how many would you need to create?
An error occurred while saving the comment Hi, rest assured that the product team at OSIsoft sees every single comment/updates that flow through this feedback website. While we're on this particular topic, I want to make sure I understand what you're asking. Notification rules are typically tied to an event frames analysis. When you say "variable", are you referring to variables that you create as part of the event frame analysis? Please confirm.
-
2 votes
An error occurred while saving the comment How would you enter the free form JSON? By hand?
-
20 votes
An error occurred while saving the comment How long do you expect these analyses to be suspended? The original implementation of a "suspended" state is based on the assumption that it's a somewhat transient condition and at some point in time, you would "correct" the suspended analyses to run appropriately. Please describe your use case.
Please provide examples of what you're trying to do. It may be already possible to do what you need. For example, you can configure a PI Point Data Reference attribute with StartTime as the time context, in which case your calculations would use the Event Frame start time. In addition, in the newer release, we support the EventFrame("StartTime") function to provide time context when you configure output-at-close.