time windowing appears to be newly broken

[contributor51]

UPDATE: this issue only occurs on the first pass of the analytic. After the first pass everything seems to be OK.

eca client version 0.9.42

1. Create a new data type representing and array of doubles.

2. Map the data type to an input channel using time window syntax "last 1 seconds" or "last 1 second @ 30 per second" (both declarations behave the same)

3. Set SystemSettings.FramesPerSecond = 1

In debug window check the values of the input. All entries of the input data array have the same value. This is not the same behavior I got from the same mappings a month ago.

DESIRED: if the time window syntax says "last 1 second" and the frame rate is 1 second, I expect the array to carry every value in the time window. If the phasor native rate is 30 per second I expect 30 values. If the phasor native rate is 60 per second I expect 60 values. First entry should be top-of-second. Last entry should be the last timestamp before top-of-second. If there is any missing data it should be filled with zero or NaN (your choice) and the metadata should indicate the measurement is bad.

[chefsteph9]

I tried to replicate the issue in v0.9.44 and could not, even on the first pass of a new analytic.

The only issue I had was that for the first couple seconds after the analytic started up, all the values in the array were the same. After that, however, the issue cleared up however and I got actual values.

Are you still having this issue? Do you have any more details on how you were able to produce this issue?
If not, I will close this issue.

Thanks,
Stephen Jenks

[contributor51]

My fault on the initial publication of the issue. I modified, i.e. updated, the issue on github to note that the "error" only occurs on the first pass. I believe it is still an issue, but not as serious as first thought. Thanks for looking into it. Your duplication of the issue is the same as mine.

…

________________________________ From: Stephen Jenks <notifications@github.com> Sent: Monday, May 15, 2017 10:58:40 AM To: GridProtectionAlliance/openECA Cc: Donnelly, Matt; Author Subject: Re: [GridProtectionAlliance/openECA] time windowing appears to be newly broken (#38) I tried to replicate the issue in v0.9.44 and could not, even on the first pass of a new analytic. The only issue I had was that for the first couple seconds after the analytic started up, all the values in the array were the same. After that, however, the issue cleared up however and I got actual values. Are you still having this issue? Do you have any more details on how you were able to produce this issue? If not, I will close this issue. Thanks, Stephen Jenks — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#38 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AFpk8H7Oios37_movllx1IYB45YmEK97ks5r6IRAgaJpZM4NZ7CI>.

[chefsteph9]

Thanks for your contribution @contributor51!

[contributor51]

Just looping back around on this issue... It is not a bug anymore. The only reason it is still an issue is because, as chefstep9 notes, "The only issue I had was that for the first couple seconds after the analytic started up, all the values in the array were the same."