Feature request: Get values from multiple tags at the same time.

[Killian-Ryan]

Hugo thanks for sharing your code. I have a feature request;

Description

tags is a list of PIPoint objects. The following code works for me;
my_single_data = PIconnect.PIData.PISeriesContainer.interpolated_values(tags[0],"2018-03-19 20:00:00","2018-03-19 20:02:00","00:00:30")

It would be nice if the function could take a list of PIPoints and return an array of data.
my_data_table = PIconnect.PIData.PISeriesContainer.interpolated_values(tags,"2018-03-19 20:00:00","2018-03-19 20:02:00","00:00:30")

[Hugovdberg]

I'll have to look into this type of call, especially in detecting whether the function was called as a method or statically.

The recommended call is as follows:

import PIconnect as PI
import pandas as pd

with PI.PIServer() as server:
    tags = server.search(query)
    pd.concat([tag.interpolated_values(start, end, interval) for tag in tags], axis = 1)

The main advantage is that the context manager (called by the with statement) also handles connecting and disconnecting from the server. Do you have a case where your call has advantages over this construction?

[Killian-Ryan]

Yes, I do use the context manager in either case, so I should clarify the intent of my question...
Instead of iterating over the tags in Python and making a call each time to OSI-PI, could we send a list of tags to OSI-PI in a single call? The reason I am asking, is because I wonder if it would return the data faster in the case of many tags.

[Hugovdberg]

Ah now I see what your pointing at. That would actually be nice if some performance could be gained there. Looking at the AF SDK documentation it seems there is the concept of a PIPointList which has methods to get both interpolated and recorded values. I will dive into this to see if I can make PIPoint.search return a list like object instead of a plain Python list which supports the PIPointList methods.

[mactone]

I'll have to look into this type of call, especially in detecting whether the function was called as a method or statically.

The recommended call is as follows:

import PIconnect as PI
import pandas as pd

with PI.PIServer() as server:
    tags = server.search(query)
    pd.concat([tag.interpolated_values(start, end, interval) for tag in tags], axis = 1)

The main advantage is that the context manager (called by the with statement) also handles connecting and disconnecting from the server. Do you have a case where your call has advantages over this construction?

Hugo,
I've tried to use PI tags as query. It fails.
I used the code as below trying to get the recorded_values of two tags.
No error also NO RESULTS came back.

I am struggling to use it due to my poor python.
could you show me a more detailed example on how to define the query?

import PIconnect as PI
import pandas as pd

start='2016/06/01 00:00:00'
end='2016/06/05 00:00:00'

query=['W13_AI-DR08.PV','W13_TI-DR06.PV']

with PI.PIServer() as server:
    tags = server.search(query)
    pd.concat([tag.recorded_values(start, end) for tag in tags], axis = `1)

[Hugovdberg]

Backreference from #526: Requesting data for multiple points in a single query should use the same timestamps for all underlying tags.

[Hugovdberg]

Backreference from #502: Possibly provide an option to simplify the returned value to a single point if only one point is requested.

[niltecedu]

Hiya,

So we have a working version of this using PIPointList and a wrapper built by us which returns search using PIPointList and its 3-4x faster for around 35 pitags.

@Hugovdberg

If needed I can share the code with you so you can integrate it?

[Hugovdberg]

@niltecedu I was actually just implementing a PIPointList wrapper as well, but please share your version so we can take the best features of both implementations!

[niltecedu]

Hi @Hugovdberg sorry for the late reply.

I have uploaded our implementation to the repo here: (dead)
Censored our server names. You can change it to your while testing.

We call the load_pidata file to load our data into a dataframe.

New_pireader contains all of the functions to retrieve the data, you might see loads of hacky code to make it work however I found that was the only way.

there's a new function in PI.py name search_list which returns a pipointlist. This is used in new_pireader to call the retrieve functions.

This was all updated to use pythonnet 3 so there might be change in other places in our piconnect. I updated the older version, this all happened in november last year when the project wasn't updated for a while :P

I don't think there is a way to call that search like yours. However we got a speed from 3x-10x while pulling the data, so it's definitely a very good addition.

(sorry had to take down the repo for sensitive data reasons)

[Hugovdberg]

Thanks for sharing, I was working on a wrapper around the PIPointList class that would allow it to be a drop in replacement for the list that's currently returned by search. That way you could either use it the old way, or call the download methods directly on the list.

[niltecedu]

@Hugovdberg

So we tried to do that however the functions called by PIPointlist are different in comparision to PIPoint, that's why we decided to have a separate wrapper for it.

[KojoBaiden]

Is there any way to loop through a list of Pi Points (over 300 in my case) from a Pi Server to determine which ones are stale or bad without having to pull their data individually and trending them?

[RafieeAshkan]

Hi
I was wondering if there has been any update on PIPointList wrapper? This would be great feature to have
Thanks

[dcoe-jkpaper]

Hi, Any update on this