Issue with pulling interpolated data from PI

[Former Member]

Hi All,

I am facing an issue for pulling interpolated values from PI by using History mode.

My requirement is to pull 5:30 am data from PI.Since data is coming to PI from DCS,so there are lots of entries for each day.

But I want 5:30 am interpolated data.

Suppose I am pulling March month data.

So in Tag Query,I did the following steps:

1) History mode

2) Start date as 1st march 5:30 am and End Date as 31st March 5:30 am and rowcount as 31.

It is not giving me 5:30 am interpolated data.only ist row and last row are coming 5:30 am.Other rows are not giving 5:30 am data.

I am facing this issue only for March month.For other months,it is working fine.

Problem is for 11th March.If I select 12th March to 31st March,it is working fine.If I select 1st march to 10th March,it is working fine.

So I would like to know is it because of any data?

Any idea why it is happening?

Thanks & Regards

Manisha

[agentry_src]

Hi Manisha,

In order to get all the interpolated data at 5:30 AM, you need to have Row Count set to 32.  The data set will include both endpoints, so you need the extra Row of data.  You may have to have a calculated value for each month as the day count changes.  Easy to do, but a fixed value will not work.  It may be that 31 worked fine for April (30 days + 1 = 31 for Row Count).

Regards, Mike

Message was edited by: Michael Appleby

[Former Member]

Thanks Mike.

I have not fixed the rowcount.I am passing rowcount dynamically based on Month by calculating the no of days.

For April Month,it is 30 as rowcount.It is giving the 5:30 am interpolated values for each day.

For March Month,I tried for 32 as well.Still it is the same issue.I am using xMII 11.5 version.

Any idea why it is happening?

Regards

Manisha

[agentry_src]

Can you post your results for History Query?  You may also be running into a problem where there is a discontinuous set of data in PI.  That can sometime cause interpolation problems.  You said that there was a problem with a single day.  I would check the raw data to see if there is a break in the data recording.  Or check with the Process Engineers for PI to see if there was a time that the equipment (either PI or the piece of process equipment which provides the data) was down.

Regards, Mike

[Former Member]

Hi Mike,

In PI for 11th March data is available.If I select SD as 10th march and ED as 11th March and rowcount as 2,it is giving correct result.

If I select SD as 10th March and ED as 12th march and rowcount as 3,it is not giving the correct result.

It is causing issue because of 11th march.If  I select 12th march to 31st march,it is giving correct result.

Data available for 11th March:

100     11-Mar-2012 04:32:00

102     11-Mar-2012 04:44:00

111     11-Mar-2012 04:56:00

116      11-Mar-2012 05:10:00

132      11-Mar-2012 05:52:00

141      11-Mar-2012 06:14:00

[Former Member]

Hi Mike,

Can you please share your email id.I shall share how the History mode is giving data for March Month.

Regards

Manisha

[agentry_src]

Looks like you are missing most of the data for March 11.  What does the data for March 10 and 12 look like? 

[Former Member]

Hi MIke,

Data for 10th March:

193     10-Mar-2012 04:20:00

203     10-Mar-2012 04:46:00

200      10-Mar-2012 04:54:00

206      10-Mar-2012 05:14:00

207       10-Mar-2012 06:14:00

Data for 12th march:

190       12-Mar-2012  04:16:00

190        12-Mar-2012 05:16:00

190       12-Mar-2012  06:16:00

[agentry_src]

Hi Manisha,

History mode provides interpolated data.  I need to see the raw data or you need to talk to the PI folks.  Why does the data only show up between 4:00 AM and 6:30 AM?

Thanks, Mike

[Former Member]

Hi Mike,

The data I have shared i.e. from PI.In PI data is there for entire day.But I showed you between 4 am to 6:30 am.Because I am pulling the interpolated data at 5:30 am.

Still I am unable to find why it is happening only for 11th March.

Regards

Manisha

[agentry_src]

Hi Manisha,

You will need to look at all the raw data from March 10th, 11th, and 12th.  If there is a break in the data, you should see a period of time with no entries or null data with datetimestamps.  Could be either depending on the source of the problem.  But since you are interpolating on 24 hours of data, you need to look at the all the raw data for the time between the points.  You cannot use the History Mode to do so as it is interpolating all the data.  You need to look at the raw data.  Use the History Event mode (I think I remember this right).

Regards, Mike

[Former Member]

Hi Mike,

I think this issue is happening due to raw data.For every year March month,the same issue is happening.

In Tag Query,I am doing SD as 2012-03-01 05:30:00 and ED as 2012-03-31 05:30:00 and rowcount as 31.So I should get 5:30 am interpolated data for each day.But i am not getting like this.In email,I have attached the result.

If I do SD as 2012-04-01 05:30:00 and ED as 2012-04-30 05:30:00 and rowcount as 30,I am getting 5:30 am interpolated data for each day.

So how to handle this in MII?

Thanks

Manisha

[agentry_src]

Hi Manisha,

Check the system to see when Daylight Savings time activates and see if that is where the problem lies. 

Regards, Mike

[Former Member]

Hi Mike,

Thank you very much.

In 2012,Daylight savings time was in 11th March in Canada.I checked for Nov 2011.Again same issue  persists.So which month daylight savings time happens,issue is for that month.It is not able to interpolate data.

I am using xMII 11.5 version.I would like to know whether this is bug in xMII 11.5 version.

If it is not a bug,then how to handle this.

Appreciate your help in this.

Thanks & Regards

Manisha

[agentry_src]

Hi Manisha,

It is not really a bug in MII.  The problem is the time for that one hour is indeterminate in your systems.  There are a few things you can try.  See if the the data can be retrieved using UTC. You can also set that value to null for the one hour twice a year when the time is indeterminate.

I suspect that you will not be able to adjust to the time unless your MII server does not adjust for DST (I am really beginning to wish we did away with it completely as archaic!). 

Regards, Mike

[Former Member]

Thanks Mike.

I did not find anything which can be taken care in MII to avoid this.

In my scenario,in transaction also we can not put any conditions to avoid this.Because based on selected SD and ED,in transaction I am calculating the no of days and passing to Tag query.

So in PI we have to take care day light savings.Currently in 11th March,after 1 am there is record for 3 am.no records for 2am.So in PI,we need to set zero for 2 am.

Please suggest me whether this should be right approach or any other approach we can follow.

Thank you very much on this.

Regards

Manisha

[Former Member]

Hi Mike,

I got a thread in SDN on how to handle daylight saving:

http://scn.sap.com/thread/701406

We need to update Java version.There is nothing to be done in MII to handle this.

Thanks & Regards

Manisha

[agentry_src]

Hi Manisha,

Thanks for posting the thread link.  This seems to come up periodically and it is also refreshing to see solutions from that far back still applicable. 

Best Regards, Mike