Deleting individual data points inside a bucket
This request comes from an issue of sometimes a multisensor in my Home Automation System sends out a crazy wild erroneous reading and it then totally throws of the charting/graph visual.
Would be nice to be able to delete single data points or a group/chunk of data points inside of a bucket for times like this. Thanks!
Real-time expressions provide the ability to filter out unwanted data points and data ranges in Tiles and Waves (http://support.initialstate.com/knowledgebase/articles/1093294-real-time-expressions-filter-out-unwanted-data).
For accounts that do not have real-time expressions, you can set the min/max y-axis for line graphs and min/max range for gauge views to prevent an erroneous data point from dominating the view in Tiles. See http://support.initialstate.com/knowledgebase/articles/871260-tiles-set-line-view-min-and-or-max-y-axis for details.
Strongly agree with "james wolf" as I just ran into the same issue with erroneous data and basically destroying the dataset. Having the ability to delete data from the source view would make this issue manageable.
james wolf commented
This would be simple if you just made the Source ( https://app.initialstate.com/#/source ) line editable fields, or even if you made lines removable by line number.
I just started paying for your service and only now realize that when an errant number comes in, it basically destroys the data set. It make all the views you have look wonky. Your solution to destroy the data in the bucket and recreate doesn't make sense because we are forced to lose the whole history. To me, as an application developer myself, it seems like deleting records my line number on the Source view would be the easiest solution for you to implement quickly and with few code changes.
Thomas Jackson commented
This would be a nice feature +3
=clip(lowBound, highBound) is what you are looking for (http://support.initialstate.com/knowledgebase/articles/506758-expressions-clip-lowbound-highbound). Saving expressions are coming very soon and will work across all apps (Tiles, Waves, Lines).
Is there a single filter expression to filter out data <0 and >8000. I've tried several but am missing it. I don't suppose there's a way to save a filtered data set so you don't have to enter the filter each time?
Tim Hodge commented
> We are working on a big update to expressions
That's fine, but actually being able to remove bad data would be way better. Instead of fudging expressions and views, just point, click, delete.
> you can set the min/max y-axis for line graphs
It's close, but not perfect: http://i.imgur.com/egpek1F.png . The graph is better but you're still displaying the bad data point above it.
For scenarios of test data, we recommend simply using a new bucket key and archiving the old bucket with the test data. This ensures that when you have clean data, or "versioned" data, you have it separated into logical buckets
Robert Read commented
Agree, would liek to delete test data from a bucket once I have it working the way I want.
Happy if the option was to just delete all data in the bucker as I suspect that is easier to implement than deleting a range.
Tony Abbey commented
I can only see how to set the min and max value of a tile - it would be good if that set the min/max for all the views of that variable. I have exactly the same problem as JTroska, so deleting or correcting the point from the data in "Save as a file" would still be useful.
This would be excellent!
Timothy Swieter commented
Sorry, typo there, I mean..."delete data in a bucket". Without removing the bucket.
Timothy Swieter commented
Is there access through the API to delete data in a box, individual events or all data?