ROSIsoft, OSIsoft PI and R - Version 2.1

Blog Post created by ernstamort on Oct 2, 2019

Like • Show 8 Likes8
Comment • 9

Hi Guys,

Just a quick update for those who were using the ROSIsoft R Package.
I developed and released this library a few years ago (check out my post here ROSIsoft, OSIsoft PI and R - Version 2 ), just as a community package for anyone interested in playing around with OSI-PI and R.
I was revisiting the project recently, and made some minor updates to add support for R version 3.5+.

The library was built with a fairly simple purpose ; create a way to access PI data directly out in R, and into a data frame.

Let me know if you have any problems with the latest built.

For a copy, bug fixes or enhancements just drop me a quick note.

Outcomes

9 Comments

matapetrol Nov 2, 2019 2:58 PM
Hi Ernst,

Thanks for this awesome package! I am using ROSIsoft in R3.4.4, and finding problems with GetPointSummaries:

GetPointSummaries(PointName = tag,StartTime = '*-10d' , EndTime = '*-1h',SummaryType = 'all',CalculationBasis ='TimeWeighted',Seconds = 60*60)

The above function works sometimes for some tags, but it is not consistent. Any idea why?

It returns this:

9999-12-31 20:00:00 NaN NaN NaN NaN NaN NaN NaN NaN 0

Did you fix this issue in the latest version? How can I get it?

Regards,

Carlos Mata
Like • Show 0 Likes0
Actions
- ernstamort @ matapetrol on Nov 4, 2019 4:36 PM
  Hi Carlos,
  
  No problem, I can look into it. Could you help me to narrow it down:
  
  Could you check the pi tags that causes the error?
  
  Does it work with Excel DataLink? What are the results?
  
  Glad you like it
  
  Holger
  Like • Show 0 Likes0
  Actions
  - matapetrol @ ernstamort on Nov 4, 2019 5:55 PM
    Hi Holger,
    
    I haven't learned Excel datalink so far :s, but it works in my VBA macro with PI SDK:
    
    For analog tags I use:
    
    Set ra = sampdata.Summaries2(dateStartPI, dateEndPI, interval, asAverage, cbTimeWeightedDiscrete)
    
    For discrete tags I use:
    
    Set rd = sampdata.FilteredSummaries(dateStartPI, dateEndPI, interval, "'" & tag & "' = ""OPEN"" or '" & tag & "' = ""TRAVEL""", asCount, cbTimeWeighted)
    
    It's getting 30K rows for each of ~1000 tags. I run the above lines to push aggregated data to a relational database to run hydraulic models and do analytics.
    
    Regards,
    
    Carlos Mata
    Like • Show 0 Likes0
    Actions
paddleguy Nov 25, 2019 3:20 PM
I cant seem to locate the ZIP file for the 3.5+ update.
Like • Show 0 Likes0
Actions
- ernstamort @ paddleguy on Nov 25, 2019 3:25 PM
  Hi, I removed the zip file from he post. Please send me and email to hamort@tqsintegration.com with a shared folder and I will upload it ....
  Like • Show 1 Like1
  Actions
  - paddleguy @ ernstamort on Nov 25, 2019 3:34 PM
    Done, thanks cant wait to start trying it out.
    Like • Show 0 Likes0
    Actions
viul Jan 10, 2020 10:03 AM
Hi,

First of all thank you for your work, it helps us a lot.

Is there a correct way to handle Digital States? The connection had a comm fail during a short period and we are getting the number of the digital state, 313 in this case, instead of a String. I found no way to distinguish the digital state values (193 to 318 are possible digital states) from an actual measurement. One possibility is to check if every tag from this connection has the same value, but I was wondering if there is a better way.

Thanks in Advance, Viktor
1 person found this helpful
Like • Show 0 Likes0
Actions
- ernstamort @ viul on Jan 13, 2020 2:18 AM
  Great question! OSIsoft data values are objects instead of value types. So there is really no analog to the digital state. I flattened it to an integer, but as you mentioned this can cause issues if you cant filter the value. I think NaN would be the most proper way of doing this?
  Like • Show 0 Likes0
  Actions
rosorio May 18, 2020 9:11 PM
Hi Ernst, thanks for the package!
I can't make it work though. I downloaded rClr 0.8.3, to make it work with newer versions of R, but even with a simple query: GetPointValue('sinusoid','* - 1h','Interpolated','Auto'), i get this:
TimeStamp Values
1 10000-01-01 NaN

Can you give me any advise?
Like • Show 0 Likes0
Actions

ROSIsoft, OSIsoft PI and R - Version 2.1

Outcomes

Related Content

Recommended Content