FAQ and Troubleshooting

Collection of various tips and tricks concerned with the dataset editor.

Function Editor

If there is a statistic function in one branch with events, then you also need a statistic function in other branches that include events.

The division of an integer or long is executed as a division with remainder. If you want to execute a division without remainder cast the dividend to double.

I do not like the way my timestamp is displayed. Use the Time > Format function in the function editor (for details in the function see: Functions of the Function Editor) to set the format you want. If you want your timestamp displayed as a number, keep in mind that we are using a UNIX timestamp in milliseconds.

How can I display the difference between two timestamps in a useful way? There is a specific data type for this in the function editor. Pick Time as a data type in the output settings of your function to display your time difference in a readable format. Of course, you can also display the difference as a number by choosing Integer/Long as a data type. Keep in mind that the timestamps are in milliseconds, so if you want your result in seconds additionally divide the time difference by 1000.

Exporting Data

You have the option to export your data in .csv format. Just click on the three dots in the top right corner of the data set editor and select "Export as CSV" to export all the data from this query up to the specified limit (the order is the same as the one displayed).

If I would export the first 5 rows of the above result I would get the following result, (S7Data) Thing;(S7Data) Timestamp;Speed DemoMachine1;1568648445276;18 DemoMachine1;1568648441555;20 DemoMachine1;1568648440670;18 DemoMachine1;1568648434326;14 DemoMachine1;1568648433015;10

As you can see, we use ; as a separator, the timestamp is exported in the UNIX timestamp format in milliseconds and the header is included in the .csv.

Where is my data?

You can not see your new dimensions in the dataset editor? Have you synchronized your dimensions? You can do so by clicking on the three dots in the top right corner and then selecting "Sync Dimensions".

Are you missing your newly defined event schema? Have you synchronized your event schema? You can do so by clicking on the three dots in the top right corner and then selecting "Sync Event Schema". If you see the schema but there is no data, then you might have a problem with the SF Edge, or the data is invalid by the definition of your schema (you can check for invalid messages in the Event Schema overview).

I did select some columns but there are no rows displayed or they are not updated.

Did you get an error message, at the bottom of your screen? If your system is using timescale DB a popular reason is the use of columns from two different event schemas. If this is the case you need to include the metadata.Timestamp column to join the two event tables together.

Some of the columns have a red dot displayed in the header, like this By hovering over the red dot you can get detailed information about the problem. Some examples are unfinished or invalid function columns, not all columns having a grouping or aggregation, the use of only columns from the metadata event.

You did not get an error message nor do you see a red dot in one of the columns? Please check your filters , maybe the restrictions on the query do not return any data.

Where is my dataset used?

Go to the overview screen of the dataset builder (for example by clicking on the dataset builder in the left navigation sidebar). In the overview screen, you can see next to every dataset. Click on it and then on show details. Now a window should pop up. There is a section called usage where you can find the information about the usage of this dataset.

How can I apply to the data itself instead of the aggregation result?

You can do this by using the Show data option. For detailed information see Additional Settings.

Why can't I see the LOCF setting?

The gap-filling feature called LOCF (Last Observation Carried Forward) is only available on environments using the Timescale database. So maybe you are running on Crate instead.

Why are there some entries at the dataset borders not filled up when using the LOCF function?

The LOCF function always fills up the gaps with the previous entry in respect of time. If the previous entry within this column is not known because it is cut off due to data limitations like a timestamp filter the previous entry is not known. In this case, it is not possible to fill up these entries.